Second IVIg Dose in Guillain-Barre syndrome patients with poor prognosis.

To determine whether a second IVIg dosage in GBS patients with a poor prognosis improve functional outcome after 4 weeks.

To investigate whether:

1. A second IVIg dosage in GBS patients with a poor prognosis improve functional outcome or muscle strength after 8, 12 and 26 weeks;

2. A second IVIg dosage in GBS patients with a poor prognosis lowers the percentage of patients needing artificial ventilation, lower the time (number of days) on respirator or time on the intensive care;

3. A second IVIg dosage in GBS patients with a poor prognosis reduces the time to hospital discharge;

4. A second IVIg dosage in GBS patients with a poor prognosis reduce the chance of secondary deterioration due to treatment-related fluctuations (TRF†);

5. Patients treated with a second IVIg dosage develop more complications possibly related to the second IVIg treatment;

6. A second IVIg dosage in GBS patients with a poor prognosis lowers the percentage of patients that die because of GBS;

7. The serum IgG increase after the first IVIg dosage is lower in patients with a poor prognosis;

8. Serum IgG increases further (and to what extent) after administration of a second IVIg dosage.

STATISTICAL ANALYSIS PLAN
Second IVIg Course in Guillain-Barré Syndrome patients with poor prognosis
(SID-GBS trial): a double-blind randomized, placebo-controlled clinical trial.
SID-GBS trial
NTR number: 2224


Erasmus MC
Rotterdam
The Netherlands

Version 1
February 18th 2019
Protocol version 12





Prepared by:
Christa Walgaard
Dept. of Neurology, Erasmus MC

Trial statistician:
Hester F. Lingsma
Dept. of Public Health, Erasmus MC

Coordinating investigator:
Pieter A. van Doorn
Dept. of Neurology, Erasmus MC

 

Introduction

The purpose of the randomized Second IVIg Course in Guillain-Barre syndrome patients with poor prognosis (SID-GBS) is to assess the safety and effect on functional outcome of a second IVIg course administered shortly after the first standard IVIg course in GBS patients with poor prognosis. The protocol was published earlier.1
Here we will summarize the statistical analysis of the data, including how missing data will be handled and which subgroup analyses will be performed. Although we list an extensive number of analyses, this does not imply that all these will be described in the main publication, because of space restrictions.
Where details differ between the protocol and the Statistical Analysis Plan (SAP), this SAP is leading.

Status of the trial

The first patient was included in the trial in February 2010. In the following years 339 patients (good and poor prognosis) were included in 53 centers in the Netherlands. In June 2018 the last patient was included and last follow-up visit took place in December 2018.

Research questions.

Primary research question

The primary objective of this study is to estimate the effect of a second IVIg course (administered 7 days after start of the first IVIg course) in GBS patients with a poor prognosis on functional outcome at 4 weeks after start of the first IVIg course.

Secondary research questions

The secondary objectives are to assess the safety of a second IVIg course with respect to the occurrence of renal dysfunction and thrombo-embolic complications, and the efficacy with regard to long-term muscle weakness, long-term functional outcome, duration of hospital and ICU admittance, duration of mechanical ventilation, and mortality.

Study Design

A double-blind, randomized placebo-controlled trial design was used in patients with poor prognosis, defined as a score of 6-12 on the mEGOS prediction model one week after start of the first IVIg course.2 Selected patients who met the inclusion and exclusion criteria and gave consent were randomly allocated (1:1) to receive a second course of IVIg (Nanogam®) in a dosage of 0,4 g/kg (=8 ml/kg) for 5 days or placebo (Albumin 4%, GPO®) in a dosage of 8 ml/kg for 5 days shortly after the first standard IVIg treatment for GBS.

Inclusion and exclusion criteria

For a complete list of inclusion and exclusion criteria we refer to paragraph 4.2 and 4.3 of the protocol version 12 on page 12-13. In short, patients of 12 years or older diagnosed with GBS according to the treating neurologist, with onset of weakness less than 2 weeks ago are in principle eligible for inclusion in the trial.

Primary and secondary outcomes

The primary outcome is the score on the GBS disability scale3 at 4 weeks after start of the first IVIg course.

Secondary outcomes are:
• GBS disability score at 8, 12 and 26 weeks.
• MRC sum score (summation of MRC scores of 6 bilateral muscles groups; m. deltoid, m. biceps, wrist extensors, m. iliopsoas, m. quadriceps femoris, m. tibialis anterior, ranging from 0-60) at 4, 8, 12 and 26 weeks.4
• ONLS score at 4, 8, 12 and 26 weeks.5
• Improvement of 1 point or more compared to baseline/randomization on the GBS disability score at 4, 8, 12 and 26 weeks, for comparison with previous studies.
• Need for artificial ventilation.
• Days on respirator.
• Length of stay on intensive care unit
• Mortality
• Length of stay in hospital
• Secondary deterioration due to treatment-related fluctuations (TRF).
• Any complications possibly related to a second IVIg course.
• Serum IgG levels at 5 different time points.

Randomization and blinding

A web-based computerized random number generation procedure from an external party (Clinical Trial Center Maastricht; CTCM) was used and treatment allocation was in a 1:1 ratio with use of block randomization and was stratified according to participating center. The local pharmacy prepared and blinded the trial medication according to a standardized protocol. Placebo (albumin) was matched to the study drug by volume (8 ml/kg) and fluid aspect (due to proteins in IVIg and albumin, both are slightly foaming liquids). As the color of IVIg can differ between batches of IVIg, the bag (ethylene vinyl acetate; EVA bag) of the trial medication was blinded using aluminum foil and opaque connecting lines were used.

Statistical Principles

Primary effect analysis

The main analysis of this trial consists of a single comparison of the primary outcome between the treatment groups. The analysis is based on the intention-to-treat principle (all patients who were randomized and the allocated trial medication was actually started will be included in an intention-to-treat analysis). The primary effect parameter is estimated with a proportional odds regression model to take the whole range of the GBS disability scale into account, and is defined as a proportional odds ratio for the effect of treatment with a 95% confidence interval6, and the corresponding p-value. A p-value of <0.05 will be considered statistically significant. Covariate adjustment 7will be used to adjust for variation in baseline prognostic variables between intervention and control group, with the following covariates:
• Age,
• Preceding diarrhea,
• MRC sum score4 at randomization.

The proportional odds (PO) model assumes that the odds ratios (the estimated effect of the treatment) is comparable for each cut-off of the outcome scale. However, it has been argued that if there is agreement that each score on the outcome scale is more favourable than a one point higher (in case of the GBS disability score) score, statistical testing of the PO assumption is redundant.8 In addition, even when the PO assumption is formally violated, the common odds ration from the proportional odds model provides an interpretable summary measure of the treatment effect over the full outcome scale and increases statistical power compared to dichotomization of the outcome scale.9

Primary effect analysis in subgroups
The primary effect parameter will be estimated in the following subgroups, using the same analysis as described above.
Heterogeneity of the treatment effect between subgroups will be tested with inclusion of an interaction term between treatment and the specific subgroup in the model.
• Age 60 or more versus less than 60 year.
• MRC sum score zero versus greater than zero at randomization.
• MRC sum score 0-12 versus greater than 12 at randomization.
• mEGOS <10 versus mEGOS 10-12 at randomization.
• Mechanical ventilation versus no mechanical ventilation.
• A-CIDP (final diagnosis) versus GBS (final diagnosis).
• A-CIDP (final diagnosis) and patients with IVIg treated TRF’s versus GBS (final diagnosis) patients without IVIg treated TRF’s.
• Demyelinating nerve conduction studies versus axonal nerve conduction studies according to the Hadden criteria.10
• Inexcitable nerve conduction studies versus all other Hadden classifications.
• 50% of the patients with the lowest delta IgG (IgG level before standard IVIg compared with IgG level at day 7-9; before start of SID) versus 50% highest delta IgG.11
• with positive campylobacter jejuni serology versus negative campylobacter jejuni serology.
• Patients with GM1 and/or GD1a antibodies versus patients without these antibodies.

Secondary effect analysis

Secondary outcomes will be will be analyzed in the same way as the primary outcome, i.e. in a regression model with covariate adjustment with the same covariates.

Ordinal secondary outcomes (GBS disability score at week 8, 12 and 26, ONLS at week 4, 8, 12 and 26) will be analyzed with proportional odds regression resulting in an adjusted odds ratio with 95% confidence interval and the corresponding p-value as effect parameter.
Binary secondary outcomes (improvement of at least 1point on the GBS disability score at week 4, 8, 12, or 26, mortality, mechanical ventilation, intensive care admission) will be analyzed with logistic regression resulting in an adjusted odds ratio with 95% confidence interval and the corresponding p-value as effect parameter.

Continuous secondary outcomes (MRC sum score at 4, 8, 12 and 26 weeks, number of days on intensive care unit, number of days on respirator) will be analyzed with linear regression resulting in an adjusted beta with 95% confidence interval and the corresponding p-value as effect parameter. If the outcomes are not normally distributed, a suitable transformation will be used.

Secondary time-to-event outcomes (time to regain independent walking (GBS disability score 0-2), time to regain the ability to run (GBS disability score 0-1), time to hospital discharge) will be analyzed with logistic regression resulting in an adjusted odds ratio with 95% confidence interval and the corresponding p-value as effect parameter.

For analysis of (S)AE’s and serum IgG levels descriptive statistics will be used.

Additional ad-hoc analyses may be conducted as deemed suitable.

Missing data and death

After complete recovery on the GBS disability score, MRC sum score and the ONLS further follow-up was not needed according to the protocol, because GBS is usually a monophasic disease. In that case, following endpoints were inserted according to last observation carried forward.

Patients who die within the study period will be assigned the worst score on all outcome measures from that time point and remain included in the analysis.
Multiple imputation will be used to account for missing values of baseline characteristics when they were used for covariate adjustment. Multiple imputation will also be used to account for missing values of secondary endpoints. Because serial measurements were performed this will result in very reliable imputations estimates. The imputation model will use age, sex, preceding infection, number of days between start of weakness and inclusion, presence of pain, and at serial time points (entry, 1, 2, 4, 8, 12, and 26 weeks) cranial nerve deficit, MRC scores of the individual muscles from the MRC sum score, GBS disability score and ONLS score. The primary endpoint will not be imputed.
 
References

1. Walgaard C, Jacobs BC, Lingsma HF, et al. Second IVIg Course in Guillain-Barre Syndrome patients with poor prognosis (SID-GBS trial): protocol for a double-blind randomized, placebo-controlled clinical trial. J Peripher Nerv Syst 2018.
2. Walgaard C, Lingsma HF, Ruts L, van Doorn PA, Steyerberg EW, Jacobs BC. Early recognition of poor prognosis in Guillain-Barre syndrome. Neurology 2011;76:968-75.
3. Hughes RA, Newsom-Davis JM, Perkin GD, Pierce JM. Controlled trial prednisolone in acute polyneuropathy. Lancet 1978;2:750-3.
4. Kleyweg RP, van der Meche FG, Schmitz PI. Interobserver agreement in the assessment of muscle strength and functional abilities in Guillain-Barre syndrome. Muscle & nerve 1991;14:1103-9.
5. Graham RC, Hughes RA. A modified peripheral neuropathy scale: the Overall Neuropathy Limitations Scale. Journal of neurology, neurosurgery, and psychiatry 2006;77:973-6.
6. Scott SC, Goldberg MS, Mayo NE. Statistical assessment of ordinal outcomes in comparative studies. J Clin Epidemiol 1997;50:45-55.
7. Roozenbeek B, Maas AI, Lingsma HF, et al. Baseline characteristics and statistical power in randomized controlled trials: selection, prognostic targeting, or covariate adjustment? Critical care medicine 2009;37:2683-90.
8. Senn S, Julious S. Measurement in clinical trials: a neglected issue for statisticians? Stat Med 2009;28:3189-209.
9. McHugh GS, Butcher I, Steyerberg EW, et al. A simulation study evaluating approaches to the analysis of ordinal outcome data in randomized controlled trials in traumatic brain injury: results from the IMPACT Project. Clin Trials 2010;7:44-57.
10. Hadden RD, Cornblath DR, Hughes RA, et al. Electrophysiological classification of Guillain-Barre syndrome: clinical associations and outcome. Plasma Exchange/Sandoglobulin Guillain-Barre Syndrome Trial Group. Annals of neurology 1998;44:780-8.
11. Kuitwaard K, de Gelder J, Tio-Gillen AP, et al. Pharmacokinetics of intravenous immunoglobulin and outcome in Guillain-Barre syndrome. Annals of neurology 2009;66:597-603.