Journal of the Peripheral Nervous System 18:321–330 (2013)
RESEARCH REPORT
A controlled trial of intravenous immunoglobulin in multifocal motor neuropathy Angelika F. Hahn1 , Said R. Beydoun2 , Victoria Lawson3 , for The IVIG in MMN Study Team† , MyungShin Oh4 , Victoria G. Empson5 , Heinz Leibl6 , Leock Y. Ngo7 , David Gelmont7 , and Carol L. Koski8 1
Department of Neurology, London Health Sciences Centre, London, Ontario, Canada; 2 Department of Neurology, University of Southern California, Los Angeles, CA, USA; 3 Department of Neurology, The Ohio State University, Columbus, OH, USA; 4 Clinical Biostatistics, Baxter Healthcare Corporation, Westlake Village, CA, USA; 5 Clinical Scientific Affairs and 6 Clinical Research, BioTherapeutics, Baxter Innovations GmbH, Vienna, Austria 7 Clinical Research, BioTherapeutics, Baxter Healthcare Corporation, Westlake Village, CA, USA and 8 Santa Fe, NM, USA
Abstract
Intravenous immunoglobulin (IVIG) has become the standard treatment for multifocal motor neuropathy (MMN) based on limited data. To critically assess the efficacy, safety, and tolerability of 10% liquid IVIG (IVIG), 44 adults with MMN were randomized 1 : 1 to either double-blind treatment of IVIG followed by placebo for 12 weeks each or the reverse. Open-label IVIG was administered for 12 weeks at the beginning and end of the study for clinical stabilization, and between double-blinded periods to prevent a carry-over effect. To avoid potential worsening, switching to open-label IVIG was permitted if deterioration occurred during blinded treatment. Mean maximal grip strength of the more affected hand declined 31.38% during placebo and increased 3.75% during IVIG (p = 0.005). In 35.7% of participants, Guy’s Neurological Disability scores for upper limbs worsened during placebo and not during IVIG, whereas the converse was true in 11.9% (p = 0.021). Sixty-nine percent (69.0%) switched prematurely from placebo to open-label IVIG and 2.4% switched from blinded to open-label IVIG (p < 0.001). One serious adverse reaction (pulmonary embolism) and 100 non-serious reactions (69 mild, 20 moderate, and 11 severe) to IVIG occurred. IVIG was effective in improving disability and muscle strength, and was safe and well tolerated in adults with MMN.
Key words: intravenous immunoglobulin, multifocal motor neuropathy, randomized placebo-controlled trial
2010). Prevalence is estimated at 0.6/100,000, with males more frequently affected at a ratio of 2.7 : 1 (Vlam et al., 2011). Clinically, MMN can mimic other motor neuropathies and neuronopathies, thereby delaying the diagnosis (Vlam et al., 2011). The hallmark of MMN is the presence of multifocal motor conduction block at locations other than common entrapment sites, with normal sensory nerve conduction in the corresponding segments (Meuth and Kleinschnitz, 2010; Eftimov and van Schaik, 2011). An early diagnosis of MMN, based on the typical clinical presentation, electrophysiological data, and additional
Introduction Multifocal motor neuropathy (MMN) is a rare pure motor neuropathy characterized by slowly progressive, asymmetric, predominantly distal weakness, commonly of the upper limbs, and may cause considerable disability (Cats et al., 2010b; Meuth and Kleinschnitz,
Address correspondence to: Carol L. Koski, MD, 1300 Hummingbird Ct, Santa Fe, NM 87501, USA. Tel: +1 505-989-3319; E-mail: [email protected] † See Appendix for complete list of Study group members. © 2013 Peripheral Nerve Society
321
Hahn et al.
Journal of the Peripheral Nervous System 18:321–330 (2013)
laboratory parameters, is an important determinant of treatment outcome and prognosis (Cats et al., 2010b). Response to immune modulation and high titers of anti-ganglioside antibodies, such as IgM anti-GM1 antibodies, support an autoimmune etiology in MMN. GM1 is enriched in paranodal myelin and in the axolemma at nodes of Ranvier of motor nerves (Willison and Yuki, 2002; Vlam et al., 2011). High titers of IgM anti-GM1 antibodies have been found in 30–88% of patients with MMN (van Schaik et al., 1995; van Asseldonk et al., 2005; Cats et al., 2010a; Galban-Horcajo et al., 2013). Although IgM GM1 titers correlate with motor axon loss and weakness, their pathogenic role is not yet established (Cats et al., 2010a; Vlam et al., 2011). The goal of treatment is to reverse the motor conduction block and limit axonal injury that leads to irreversible functional impairment (van Asseldonk et al., 2006; Cats et al., 2010b) which underscores the need for an early diagnosis and treatment (Cats et al., 2010b). Therapeutic options for MMN are limited (van Schaik et al., 2005) as patients do not respond to corticosteroids or plasma exchange and can worsen with these treatments (Donaghy et al., 1994; Carpo et al., 1998). Uncontrolled studies suggest that cyclophosphamide may be effective, but long-term treatment is limited by significant toxicity (Umapathi et al., 2009; Joint Task Force of the EFNS and the PNS, 2010). Immunomodulatory treatment with polyclonal human immunoglobulin is currently the only safe and effective intervention for MMN patients (Meuth and Kleinschnitz, 2010; Eftimov and van Schaik, 2011). Intravenous immunoglobulin (IVIG) is a first-line therapy based on a meta-analysis of four randomized double-blind placebo-controlled trials (Azulay et al., 1994; van den Berg et al., 1995; Federico et al., 2000; ´ Leger et al., 2001) involving a total of 34 patients; 78% of study subjects had a significant short term improvement in strength following IVIG treatment compared to 4% following placebo (van Schaik et al., 2005; Joint Task Force of the EFNS and the PNS, 2010). The meta-analysis, however, did not show a significant improvement in disability and identified a need for further studies (van Schaik et al., 2005). The present trial was conducted to evaluate the efficacy of 10% liquid IVIG (thereafter referred to as IVIG) in improving not only strength but also limiting disability in 44 adults with MMN.
Figure 1. Overview of study design.
the sequence of double-blinded treatment with IVIG followed by placebo, or the alternate order. Each blinded period was to last 12 weeks. Additionally, three 12-week periods of open-label IVIG treatment were scheduled at the beginning and end of the study for clinical stabilization, and between the blinded periods to prevent a carry-over effect. To limit potential motor axonal damage associated with IVIG withdrawal, subjects were permitted to switch from blinded treatment to the next open-label IVIG treatment period without breaking the blind (referred as ‘‘accelerated switch’’) as soon as deterioration interfered significantly with daily activities or if grip strength declined by ≥50% in the more affected hand. If baseline function was not regained after the switch to open-label IVIG, a dose increase and/or administration of IVIG at a shortened dosing interval was permitted (‘‘rescue’’ treatment) and no further blinded treatment was administered. An independent committee monitored subjects’ safety throughout the study. The study was conducted according to Good Clinical Practice and registered on ClinicalTrials.gov (NCT00666263). The protocol was approved by the local institutional ethics committees. Subjects provided written informed consent.
Materials and Methods Study design
Participants
This was a phase III, randomized withdrawal, double-blind, placebo-controlled, cross-over study (Fig. 1). Subjects were randomized 1 : 1 to determine
Seventeen medical centers in the United States, Canada, and Denmark enrolled 44 adults (aged ≥18 years) diagnosed with probable or definite MMN 322
Hahn et al.
Journal of the Peripheral Nervous System 18:321–330 (2013)
according to American Association of Electrodiagnostic Medicine criteria (Olney et al., 2003). Subjects were eligible if they were on a stable regimen of IVIG for ≥3 months at a dose of 0.4–2.0 g per kg body weight (BW) every 2–5 weeks.
period. 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. Assessors were blinded to previous assessment results. Two predetermined co-primary efficacy endpoints included maximal grip strength in the more affected hand measured with a DynEx digital dynamometer, and disability as determined by the upper limb portion of Guy’s Neurological Disability Score (GNDS) ranging from 0 (no upper limb problem) to 5 (unable to use either arm for any purposeful movements) (Sharrack and Hughes, 1999). Secondary efficacy outcomes were: requirement for accelerated switch, decline of ≥30% in grip strength in the more and less affected hands, maximal grip strength in the less affected hand, overall disability sum score (ODSS) (Merkies et al., 2002), time required for the 9-hole peg board test with the dominant and nondominant hand, patient global impression of change score (PGICS) and a visual analogue scale (VAS) of disability. A post hoc endpoint was defined to illustrate a clinically relevant worsening of disability: the proportion of subjects experiencing a change of at least 1 grade in the ODSS. Safety and tolerability were assessed by adverse events (AEs; excluding symptoms related to withdrawal of IVIG during placebo administration), adverse reactions (ARs; those AEs that were considered by the investigator to be related to a study product), and infusions that required a reduced flow rate, or were interrupted or stopped. Endpoints are described in detail in Appendix S1.
Treatment GAMMAGARD LIQUID/KIOVIG (Baxter Healthcare Corporation, Westlake Village, CA, USA) is a 10% liquid IgG preparation with osmolality similar to physiologic levels, stabilized with glycine, and formulated without added sugars, sodium, or preservatives. The placebo was 0.25% human albumin prepared with normal saline from BUMINATE 25%, or Human Albumin 200 g/l Baxter Solution for Infusion (both Baxter Healthcare Corporation), whichever was licensed locally. Pre-study IVIG dosing levels were maintained, as were treatment intervals in subjects previously treated every 2–4 weeks. Those treated every 5 weeks pre-study were changed to a 4-week interval at an equivalent dose/kg BW per week. Each treatment was divided over ≤5 consecutive days. Doubleblinded infusions of IVIG or placebo were volume matched to open-label IVIG administered prior to randomization. Infusions were administered in rates up to a maximum of 5.0 ml/kg BW/h or according to institutional protocols. Rate reductions and reasons for reductions were documented. Previously prescribed prophylactic medications were maintained throughout the study. Following the initial infusion, subjects in the United States and Canada had the option of home infusion (given by a nurse), infusion at a center or at the study site. One subject who had been receiving mycophenolate mofetil at a constant dose for 3 months prior to enrollment continued on this treatment during the study.
Statistical methods Methods for statistical analysis were pre-defined in the study protocol and are outlined in detail in Appendix S1. The main analysis included all randomized subjects (intent-to-treat, N = 44) according to product allocated not product received. In addition, an analysis was performed on all subjects who completed both blinded periods (even if an accelerated switch occurred) provided that baseline assessments for both blinded periods were available (per-protocol; N = 42). Statistical results are provided for both analyses.
Randomization and masking Subjects were randomized 1 : 1 to IVIG crossing over to placebo or vice-versa. 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. All subjects, investigators and the sponsor remained blinded during the entire study.
Analyses of co-primary and secondary endpoints For grip strength (mean of three trials) in the more affected hand, the relative change (%) from the blinded periods’ baseline (defined as the most recent observation from the previous open-label IVIG treatment period) to the end of the blinded period (1 or 2) was analyzed using a fixed effects ANOVA model reflecting the cross-over study design, with factors for
Outcomes Results of efficacy assessments performed in the last treatment cycle of each blinded period were compared to the most recent corresponding observations on open-label IVIG, which served as a baseline value for the following blinded treatment 323
Hahn et al.
Journal of the Peripheral Nervous System 18:321–330 (2013)
Table 1. Demographic characteristics of subjects.
sequence (placebo followed by IVIG or IVIG followed by placebo), subject nested within sequence, period (1 or 2), treatment (IVIG or placebo), and baseline for the relative change. The contrast to be tested was the treatment effect of IVIG vs. placebo. The ANOVA was repeated using multiple imputations to account for missing data. Secondary outcomes representing continuous variables (relative changes in grip strength in the more/less affected hand, time for peg board test, patients’ self assessment by VAS, safety data) were analyzed similarly. GNDS for the upper limbs was analyzed as a binary variable indicating whether the score of a subject deteriorated from the blinded periods’ baseline (1 or 2) to the last cycle assessment at the end of the blinded periods (1 or 2). The cross-over design was analyzed by McNemar’s test on the discordant pairs, that is, those subjects who deteriorated in one, but not in the other blinded period. This was repeated as a sensitivity analysis using the worst imputation technique to account for missing data. Secondary outcomes representing binary variables (proportion of subjects experiencing a decline of at least 30% in grip strength in the more affected hand, proportion of subjects accelerated forward into the next stabilization phase, and proportion of subjects with worsening by one degree in the ODSS) were analyzed similarly.
Number of subjects (%)
Parameter
Category
Gender
Male Female White Black or African American Hispanic
Ethnicity
Parameter Age at consent (years) Monthly dose (g/kg/month)
32 (72.7) 12 (27.3) 41 (93.2) 2 (4.5) 1 (2.3)
Statistics
Number of subjects (%)
N Mean (SD) N Mean (SD)
44 51.64 (10.25) 44 1.17 (0.46)
(range: 13–91) for IVIG and 28 days (range: 7–86) for placebo.
Efficacy Treatment with IVIG increased the maximal grip strength of the more affected hand by a mean of 3.75% compared to a significant decrease by a mean 31.38% during placebo (p = 0.005; Table 2). The number of subjects whose disability worsened as determined by the upper limbs section of the GNDS, was also significantly different for IVIG and placebo: 15 of the 42 evaluable subjects (35.7%) deteriorated only during the placebo period, while five subjects (11.9%) deteriorated only while receiving IVIG (p = 0.021). Two subjects (4.8%) deteriorated during both periods and the remaining 20 subjects (47.6%) showed no change in GNDS scores (Table 3). The proportion of subjects who increased by at least one grade in the ODSS, that is, who showed a clinically relevant worsening of their disability, during placebo alone (13 of 42; 31.0%) was significantly higher than the proportion (3 of 42; 7.1%) that worsened during blinded IVIG treatment alone (p = 0.011; Table 4). Worsening of disability was also reflected in the PGICS (Tables 5 and S1). Median PGICS following placebo was 6.0, which indicated that the subjects judged their condition to be much worse (assessed on a scale of 1–7), whereas the median PGICS after IVIG treatment was 4.0, which indicated a maintained level of function. An accelerated switch to open-label IVIG occurred in 29 (69.0%) subjects during placebo (Table 3). In contrast, only one subject (2.4%) switched to open-label treatment while receiving blinded IVIG (p < 0.001). Of the 12 subjects (28.6%) who had not requested an ‘‘accelerated switch’’ from placebo to open-label IVIG, three had worsened in grip strength
Results Subject demographics, disposition, and exposure Subjects were aged from 31 to 72 years; 27.3% were female. Demographic characteristics are described in Table 1. The subjects had been receiving IVIG for a median of 6.0 months before enrollment (range: 2.8–184.6 months). Of the 44 eligible and randomized subjects, all but three completed the study. Two subjects discontinued while receiving IVIG due to events that were classified as AEs by the attending investigator: in one subject, muscular weakness; in the other, decreased range of joint motion. One additional subject withdrew as she moved away from the study site after having completed both blinded phases (Fig. 2). The study ran from August 2008 to August 2011. All 44 randomized subjects were included in the intent-to-treat analyses of efficacy and safety, 42 of these had available data for both blinded periods. The median total duration of IVIG treatment per subject was 335 days with a median monthly dose of 1.2 g/kg BW. During the double-blinded cross-over periods, the median treatment duration was 84 days 324
Hahn et al.
Journal of the Peripheral Nervous System 18:321–330 (2013)
Figure 2. Flow chart of subject disposition.
to a degree that would have qualified them for a switch (which they had not requested); four showed objective and subjective worsening below the threshold for an ‘‘accelerated switch’’; two showed no change in all measured parameters during both blinded treatments; and three showed either a mild improvement with placebo or a mild decline with blinded IVIG, all occurring in the first blinded cross-over period; likely representing placebo reactions. Individual patients did not always regain baseline motor function immediately after an ‘‘accelerated switch’’ and some required multiple IVIG
treatments to fully recover. Therefore, the protocol was amended to allow adjustments of the dose and/or dosing interval in such patients (‘‘rescue’’ treatment), which was subsequently required in three subjects. Dominant and non-dominant hand dexterity decreased substantially during placebo administration as indicated by the 9-hole peg test (Table 2). At the end of the placebo period, subjects required a mean of 16.73% longer to complete the test with the dominant hand and a mean of 32.78% longer with the nondominant hand compared to baseline. By contrast, 325
Hahn et al.
Journal of the Peripheral Nervous System 18:321–330 (2013)
Table 2. Percent changes in continuous efficacy variables during double-blinded IVIG and placebo. Percent (%) change compared to baseline* (SD) Endpoint
IVIG
Maximal grip strength of the more affected hand (co-primary endpoint) Maximal grip strength of the less affected hand Time required for 9-hole peg board test with the dominant hand Time required for 9-hole peg board test with the non-dominant hand Score on visual analogue scale†
ANOVA (one-sided)
Placebo
p-value
3.75 (9.09)
−31.38 (9.32)
p = 0.005
8.57 (6.40) 1.16 (3.04)
−23.96 (6.56) 16.73 (3.12)
p < 0.001 p < 0.001
6.67 (4.44)
32.78 (4.55)
p < 0.001
73.33 (94.60)
289.93 (96.99)
p = 0.059
IVIG, intravenous immunoglobulin. *Percent changes are the least-squares mean of changes across all subjects expressed as a percentage of the baseline value. Baseline was defined as the last assessment prior to the blinded period. †Higher visual analogue scale scores represent more severe disability.
Table 3. Non-continuous efficacy variables during double-blinded IVIG and placebo periods. Number of subjects (%)(N = 42) Endpoint met Endpoint met Endpoint during Endpoint during not met McNemar’s test (placebo placebo met during both during either only versus IVIG only) only IVIG only treatments treatment p-value
Endpoint Deteriorated according to the upper limbs section of the GNDS (co-primary endpoint) Required accelerated switch from blinded treatment to open-label IVIG Showed a decline of ≥30% in grip strength in the more affected hand Showed a decline of ≥30% in grip strength in the less affected hand
15 (35.7)
5 (11.9)
2 (4.8)
20 (47.6)
0.021
29 (69.0)
1 (2.4)
0 (0.0)
12 (28.6)
RESEARCH REPORT
A controlled trial of intravenous immunoglobulin in multifocal motor neuropathy Angelika F. Hahn1 , Said R. Beydoun2 , Victoria Lawson3 , for The IVIG in MMN Study Team† , MyungShin Oh4 , Victoria G. Empson5 , Heinz Leibl6 , Leock Y. Ngo7 , David Gelmont7 , and Carol L. Koski8 1
Department of Neurology, London Health Sciences Centre, London, Ontario, Canada; 2 Department of Neurology, University of Southern California, Los Angeles, CA, USA; 3 Department of Neurology, The Ohio State University, Columbus, OH, USA; 4 Clinical Biostatistics, Baxter Healthcare Corporation, Westlake Village, CA, USA; 5 Clinical Scientific Affairs and 6 Clinical Research, BioTherapeutics, Baxter Innovations GmbH, Vienna, Austria 7 Clinical Research, BioTherapeutics, Baxter Healthcare Corporation, Westlake Village, CA, USA and 8 Santa Fe, NM, USA
Abstract
Intravenous immunoglobulin (IVIG) has become the standard treatment for multifocal motor neuropathy (MMN) based on limited data. To critically assess the efficacy, safety, and tolerability of 10% liquid IVIG (IVIG), 44 adults with MMN were randomized 1 : 1 to either double-blind treatment of IVIG followed by placebo for 12 weeks each or the reverse. Open-label IVIG was administered for 12 weeks at the beginning and end of the study for clinical stabilization, and between double-blinded periods to prevent a carry-over effect. To avoid potential worsening, switching to open-label IVIG was permitted if deterioration occurred during blinded treatment. Mean maximal grip strength of the more affected hand declined 31.38% during placebo and increased 3.75% during IVIG (p = 0.005). In 35.7% of participants, Guy’s Neurological Disability scores for upper limbs worsened during placebo and not during IVIG, whereas the converse was true in 11.9% (p = 0.021). Sixty-nine percent (69.0%) switched prematurely from placebo to open-label IVIG and 2.4% switched from blinded to open-label IVIG (p < 0.001). One serious adverse reaction (pulmonary embolism) and 100 non-serious reactions (69 mild, 20 moderate, and 11 severe) to IVIG occurred. IVIG was effective in improving disability and muscle strength, and was safe and well tolerated in adults with MMN.
Key words: intravenous immunoglobulin, multifocal motor neuropathy, randomized placebo-controlled trial
2010). Prevalence is estimated at 0.6/100,000, with males more frequently affected at a ratio of 2.7 : 1 (Vlam et al., 2011). Clinically, MMN can mimic other motor neuropathies and neuronopathies, thereby delaying the diagnosis (Vlam et al., 2011). The hallmark of MMN is the presence of multifocal motor conduction block at locations other than common entrapment sites, with normal sensory nerve conduction in the corresponding segments (Meuth and Kleinschnitz, 2010; Eftimov and van Schaik, 2011). An early diagnosis of MMN, based on the typical clinical presentation, electrophysiological data, and additional
Introduction Multifocal motor neuropathy (MMN) is a rare pure motor neuropathy characterized by slowly progressive, asymmetric, predominantly distal weakness, commonly of the upper limbs, and may cause considerable disability (Cats et al., 2010b; Meuth and Kleinschnitz,
Address correspondence to: Carol L. Koski, MD, 1300 Hummingbird Ct, Santa Fe, NM 87501, USA. Tel: +1 505-989-3319; E-mail: [email protected] † See Appendix for complete list of Study group members. © 2013 Peripheral Nerve Society
321
Hahn et al.
Journal of the Peripheral Nervous System 18:321–330 (2013)
laboratory parameters, is an important determinant of treatment outcome and prognosis (Cats et al., 2010b). Response to immune modulation and high titers of anti-ganglioside antibodies, such as IgM anti-GM1 antibodies, support an autoimmune etiology in MMN. GM1 is enriched in paranodal myelin and in the axolemma at nodes of Ranvier of motor nerves (Willison and Yuki, 2002; Vlam et al., 2011). High titers of IgM anti-GM1 antibodies have been found in 30–88% of patients with MMN (van Schaik et al., 1995; van Asseldonk et al., 2005; Cats et al., 2010a; Galban-Horcajo et al., 2013). Although IgM GM1 titers correlate with motor axon loss and weakness, their pathogenic role is not yet established (Cats et al., 2010a; Vlam et al., 2011). The goal of treatment is to reverse the motor conduction block and limit axonal injury that leads to irreversible functional impairment (van Asseldonk et al., 2006; Cats et al., 2010b) which underscores the need for an early diagnosis and treatment (Cats et al., 2010b). Therapeutic options for MMN are limited (van Schaik et al., 2005) as patients do not respond to corticosteroids or plasma exchange and can worsen with these treatments (Donaghy et al., 1994; Carpo et al., 1998). Uncontrolled studies suggest that cyclophosphamide may be effective, but long-term treatment is limited by significant toxicity (Umapathi et al., 2009; Joint Task Force of the EFNS and the PNS, 2010). Immunomodulatory treatment with polyclonal human immunoglobulin is currently the only safe and effective intervention for MMN patients (Meuth and Kleinschnitz, 2010; Eftimov and van Schaik, 2011). Intravenous immunoglobulin (IVIG) is a first-line therapy based on a meta-analysis of four randomized double-blind placebo-controlled trials (Azulay et al., 1994; van den Berg et al., 1995; Federico et al., 2000; ´ Leger et al., 2001) involving a total of 34 patients; 78% of study subjects had a significant short term improvement in strength following IVIG treatment compared to 4% following placebo (van Schaik et al., 2005; Joint Task Force of the EFNS and the PNS, 2010). The meta-analysis, however, did not show a significant improvement in disability and identified a need for further studies (van Schaik et al., 2005). The present trial was conducted to evaluate the efficacy of 10% liquid IVIG (thereafter referred to as IVIG) in improving not only strength but also limiting disability in 44 adults with MMN.
Figure 1. Overview of study design.
the sequence of double-blinded treatment with IVIG followed by placebo, or the alternate order. Each blinded period was to last 12 weeks. Additionally, three 12-week periods of open-label IVIG treatment were scheduled at the beginning and end of the study for clinical stabilization, and between the blinded periods to prevent a carry-over effect. To limit potential motor axonal damage associated with IVIG withdrawal, subjects were permitted to switch from blinded treatment to the next open-label IVIG treatment period without breaking the blind (referred as ‘‘accelerated switch’’) as soon as deterioration interfered significantly with daily activities or if grip strength declined by ≥50% in the more affected hand. If baseline function was not regained after the switch to open-label IVIG, a dose increase and/or administration of IVIG at a shortened dosing interval was permitted (‘‘rescue’’ treatment) and no further blinded treatment was administered. An independent committee monitored subjects’ safety throughout the study. The study was conducted according to Good Clinical Practice and registered on ClinicalTrials.gov (NCT00666263). The protocol was approved by the local institutional ethics committees. Subjects provided written informed consent.
Materials and Methods Study design
Participants
This was a phase III, randomized withdrawal, double-blind, placebo-controlled, cross-over study (Fig. 1). Subjects were randomized 1 : 1 to determine
Seventeen medical centers in the United States, Canada, and Denmark enrolled 44 adults (aged ≥18 years) diagnosed with probable or definite MMN 322
Hahn et al.
Journal of the Peripheral Nervous System 18:321–330 (2013)
according to American Association of Electrodiagnostic Medicine criteria (Olney et al., 2003). Subjects were eligible if they were on a stable regimen of IVIG for ≥3 months at a dose of 0.4–2.0 g per kg body weight (BW) every 2–5 weeks.
period. 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. Assessors were blinded to previous assessment results. Two predetermined co-primary efficacy endpoints included maximal grip strength in the more affected hand measured with a DynEx digital dynamometer, and disability as determined by the upper limb portion of Guy’s Neurological Disability Score (GNDS) ranging from 0 (no upper limb problem) to 5 (unable to use either arm for any purposeful movements) (Sharrack and Hughes, 1999). Secondary efficacy outcomes were: requirement for accelerated switch, decline of ≥30% in grip strength in the more and less affected hands, maximal grip strength in the less affected hand, overall disability sum score (ODSS) (Merkies et al., 2002), time required for the 9-hole peg board test with the dominant and nondominant hand, patient global impression of change score (PGICS) and a visual analogue scale (VAS) of disability. A post hoc endpoint was defined to illustrate a clinically relevant worsening of disability: the proportion of subjects experiencing a change of at least 1 grade in the ODSS. Safety and tolerability were assessed by adverse events (AEs; excluding symptoms related to withdrawal of IVIG during placebo administration), adverse reactions (ARs; those AEs that were considered by the investigator to be related to a study product), and infusions that required a reduced flow rate, or were interrupted or stopped. Endpoints are described in detail in Appendix S1.
Treatment GAMMAGARD LIQUID/KIOVIG (Baxter Healthcare Corporation, Westlake Village, CA, USA) is a 10% liquid IgG preparation with osmolality similar to physiologic levels, stabilized with glycine, and formulated without added sugars, sodium, or preservatives. The placebo was 0.25% human albumin prepared with normal saline from BUMINATE 25%, or Human Albumin 200 g/l Baxter Solution for Infusion (both Baxter Healthcare Corporation), whichever was licensed locally. Pre-study IVIG dosing levels were maintained, as were treatment intervals in subjects previously treated every 2–4 weeks. Those treated every 5 weeks pre-study were changed to a 4-week interval at an equivalent dose/kg BW per week. Each treatment was divided over ≤5 consecutive days. Doubleblinded infusions of IVIG or placebo were volume matched to open-label IVIG administered prior to randomization. Infusions were administered in rates up to a maximum of 5.0 ml/kg BW/h or according to institutional protocols. Rate reductions and reasons for reductions were documented. Previously prescribed prophylactic medications were maintained throughout the study. Following the initial infusion, subjects in the United States and Canada had the option of home infusion (given by a nurse), infusion at a center or at the study site. One subject who had been receiving mycophenolate mofetil at a constant dose for 3 months prior to enrollment continued on this treatment during the study.
Statistical methods Methods for statistical analysis were pre-defined in the study protocol and are outlined in detail in Appendix S1. The main analysis included all randomized subjects (intent-to-treat, N = 44) according to product allocated not product received. In addition, an analysis was performed on all subjects who completed both blinded periods (even if an accelerated switch occurred) provided that baseline assessments for both blinded periods were available (per-protocol; N = 42). Statistical results are provided for both analyses.
Randomization and masking Subjects were randomized 1 : 1 to IVIG crossing over to placebo or vice-versa. 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. All subjects, investigators and the sponsor remained blinded during the entire study.
Analyses of co-primary and secondary endpoints For grip strength (mean of three trials) in the more affected hand, the relative change (%) from the blinded periods’ baseline (defined as the most recent observation from the previous open-label IVIG treatment period) to the end of the blinded period (1 or 2) was analyzed using a fixed effects ANOVA model reflecting the cross-over study design, with factors for
Outcomes Results of efficacy assessments performed in the last treatment cycle of each blinded period were compared to the most recent corresponding observations on open-label IVIG, which served as a baseline value for the following blinded treatment 323
Hahn et al.
Journal of the Peripheral Nervous System 18:321–330 (2013)
Table 1. Demographic characteristics of subjects.
sequence (placebo followed by IVIG or IVIG followed by placebo), subject nested within sequence, period (1 or 2), treatment (IVIG or placebo), and baseline for the relative change. The contrast to be tested was the treatment effect of IVIG vs. placebo. The ANOVA was repeated using multiple imputations to account for missing data. Secondary outcomes representing continuous variables (relative changes in grip strength in the more/less affected hand, time for peg board test, patients’ self assessment by VAS, safety data) were analyzed similarly. GNDS for the upper limbs was analyzed as a binary variable indicating whether the score of a subject deteriorated from the blinded periods’ baseline (1 or 2) to the last cycle assessment at the end of the blinded periods (1 or 2). The cross-over design was analyzed by McNemar’s test on the discordant pairs, that is, those subjects who deteriorated in one, but not in the other blinded period. This was repeated as a sensitivity analysis using the worst imputation technique to account for missing data. Secondary outcomes representing binary variables (proportion of subjects experiencing a decline of at least 30% in grip strength in the more affected hand, proportion of subjects accelerated forward into the next stabilization phase, and proportion of subjects with worsening by one degree in the ODSS) were analyzed similarly.
Number of subjects (%)
Parameter
Category
Gender
Male Female White Black or African American Hispanic
Ethnicity
Parameter Age at consent (years) Monthly dose (g/kg/month)
32 (72.7) 12 (27.3) 41 (93.2) 2 (4.5) 1 (2.3)
Statistics
Number of subjects (%)
N Mean (SD) N Mean (SD)
44 51.64 (10.25) 44 1.17 (0.46)
(range: 13–91) for IVIG and 28 days (range: 7–86) for placebo.
Efficacy Treatment with IVIG increased the maximal grip strength of the more affected hand by a mean of 3.75% compared to a significant decrease by a mean 31.38% during placebo (p = 0.005; Table 2). The number of subjects whose disability worsened as determined by the upper limbs section of the GNDS, was also significantly different for IVIG and placebo: 15 of the 42 evaluable subjects (35.7%) deteriorated only during the placebo period, while five subjects (11.9%) deteriorated only while receiving IVIG (p = 0.021). Two subjects (4.8%) deteriorated during both periods and the remaining 20 subjects (47.6%) showed no change in GNDS scores (Table 3). The proportion of subjects who increased by at least one grade in the ODSS, that is, who showed a clinically relevant worsening of their disability, during placebo alone (13 of 42; 31.0%) was significantly higher than the proportion (3 of 42; 7.1%) that worsened during blinded IVIG treatment alone (p = 0.011; Table 4). Worsening of disability was also reflected in the PGICS (Tables 5 and S1). Median PGICS following placebo was 6.0, which indicated that the subjects judged their condition to be much worse (assessed on a scale of 1–7), whereas the median PGICS after IVIG treatment was 4.0, which indicated a maintained level of function. An accelerated switch to open-label IVIG occurred in 29 (69.0%) subjects during placebo (Table 3). In contrast, only one subject (2.4%) switched to open-label treatment while receiving blinded IVIG (p < 0.001). Of the 12 subjects (28.6%) who had not requested an ‘‘accelerated switch’’ from placebo to open-label IVIG, three had worsened in grip strength
Results Subject demographics, disposition, and exposure Subjects were aged from 31 to 72 years; 27.3% were female. Demographic characteristics are described in Table 1. The subjects had been receiving IVIG for a median of 6.0 months before enrollment (range: 2.8–184.6 months). Of the 44 eligible and randomized subjects, all but three completed the study. Two subjects discontinued while receiving IVIG due to events that were classified as AEs by the attending investigator: in one subject, muscular weakness; in the other, decreased range of joint motion. One additional subject withdrew as she moved away from the study site after having completed both blinded phases (Fig. 2). The study ran from August 2008 to August 2011. All 44 randomized subjects were included in the intent-to-treat analyses of efficacy and safety, 42 of these had available data for both blinded periods. The median total duration of IVIG treatment per subject was 335 days with a median monthly dose of 1.2 g/kg BW. During the double-blinded cross-over periods, the median treatment duration was 84 days 324
Hahn et al.
Journal of the Peripheral Nervous System 18:321–330 (2013)
Figure 2. Flow chart of subject disposition.
to a degree that would have qualified them for a switch (which they had not requested); four showed objective and subjective worsening below the threshold for an ‘‘accelerated switch’’; two showed no change in all measured parameters during both blinded treatments; and three showed either a mild improvement with placebo or a mild decline with blinded IVIG, all occurring in the first blinded cross-over period; likely representing placebo reactions. Individual patients did not always regain baseline motor function immediately after an ‘‘accelerated switch’’ and some required multiple IVIG
treatments to fully recover. Therefore, the protocol was amended to allow adjustments of the dose and/or dosing interval in such patients (‘‘rescue’’ treatment), which was subsequently required in three subjects. Dominant and non-dominant hand dexterity decreased substantially during placebo administration as indicated by the 9-hole peg test (Table 2). At the end of the placebo period, subjects required a mean of 16.73% longer to complete the test with the dominant hand and a mean of 32.78% longer with the nondominant hand compared to baseline. By contrast, 325
Hahn et al.
Journal of the Peripheral Nervous System 18:321–330 (2013)
Table 2. Percent changes in continuous efficacy variables during double-blinded IVIG and placebo. Percent (%) change compared to baseline* (SD) Endpoint
IVIG
Maximal grip strength of the more affected hand (co-primary endpoint) Maximal grip strength of the less affected hand Time required for 9-hole peg board test with the dominant hand Time required for 9-hole peg board test with the non-dominant hand Score on visual analogue scale†
ANOVA (one-sided)
Placebo
p-value
3.75 (9.09)
−31.38 (9.32)
p = 0.005
8.57 (6.40) 1.16 (3.04)
−23.96 (6.56) 16.73 (3.12)
p < 0.001 p < 0.001
6.67 (4.44)
32.78 (4.55)
p < 0.001
73.33 (94.60)
289.93 (96.99)
p = 0.059
IVIG, intravenous immunoglobulin. *Percent changes are the least-squares mean of changes across all subjects expressed as a percentage of the baseline value. Baseline was defined as the last assessment prior to the blinded period. †Higher visual analogue scale scores represent more severe disability.
Table 3. Non-continuous efficacy variables during double-blinded IVIG and placebo periods. Number of subjects (%)(N = 42) Endpoint met Endpoint met Endpoint during Endpoint during not met McNemar’s test (placebo placebo met during both during either only versus IVIG only) only IVIG only treatments treatment p-value
Endpoint Deteriorated according to the upper limbs section of the GNDS (co-primary endpoint) Required accelerated switch from blinded treatment to open-label IVIG Showed a decline of ≥30% in grip strength in the more affected hand Showed a decline of ≥30% in grip strength in the less affected hand
15 (35.7)
5 (11.9)
2 (4.8)
20 (47.6)
0.021
29 (69.0)
1 (2.4)
0 (0.0)
12 (28.6)
