Supplementary material- 1 - Zipp et al.

Blockade of chemokine signaling in MS patientsMethods

Appendix (E) A-1. Methods

Study design

Twenty centers in Germany, Finland, Sweden, the Netherlands, and Belgium screened 209 patients and enrolled 105 patients from June 2003 until May 2004. The protocol was reviewed and approved by ethics committees. Before providing informed consent according the Declaration of Helsinki, all patients were advised of the alternative effective therapies available to them. RRMS patients were screened, fulfilling the McDonald criteria with an Expanded Disability Status Scale [EDSS] 0-5.5, age 18-55, with at least one documented relapse, and with at least one gadolinium (Gd)-enhancing lesion as well as at least three lesions on the T2-weighted scans at screening MRI.They were ineligible if they had received immunomodulatory or immunosuppressive treatment within three months or corticosteroids and adrenocorticotrophin treatment less than one month prior to study entry. Four weeks after obtaining an eligible MRI, patients were randomly assigned to 600 mg BX 471 or placebo, orally administered three times daily over 16 weeks. Each patient made 9 regular visits to the study site with a total of 7 MRI examinations: screening MRI, baseline scan, scans at week 4, 8, and 16 of the treatment as well as at week 4 after treatment (week 20 of the study).

Efficacy endpoints

The primary endpoint was the cumulative number of newly active lesions (new Gd-enhancing lesions on T1-weighted scans and new lesions on T2-weighted scans) after 16 weeks. Secondary outcome variables included the volume of hyperintense lesions on T2-weighted scans (“T2-lesion load”) at week 16, the proportion of “active” scans per patient (fraction of active post-baseline scans among those performed post baseline for a patient), and “active patients” (throughout the entire study), defined as those with at least one “active” post-baseline scan. Other planned targets were immunological parameters and expression of chemokine receptors on lymphocyte and monocyte measured by flow cytometry. (CD14, CD16, CD11b, CD4, CD8, CD3, CD56, CD25, HLA-DR, CD40, CD86, CCR2, CCR5, IFN-γ, TNF-). For fresh blood immunological studies, patients were included from seven centers. Change in brain volume from baseline to week 20 was measured post-hoc.

Safety and tolerability

The safety and tolerability was assessed by physical examination (including relapse assessment), vital signs, and MRI scans at each visit. EDSS scoring and ECG were performed at screening, baseline, and at end of study. Study safety was monitored by the independent data monitoring board. Blood samples for the analysis of BX 471 plasma concentrations were taken at every visit after baseline up to visit week 16.

Procedures

MRI scans were acquired in 20 sites (16 scanners at 1.5 T and 4 at 1.0 T) following a predetermined MRI protocol with fixed acquisition parameters and careful check for a correct image repositioning. All MRI analysis was performed blinded at the Neuroimaging Research Unit, Scientific Institute and University Ospedale san raffaele, Milan, Italy. The following sequences were obtained: a) conventional or fast spin-echo (SE) dual echo (TR 2200-2800, TE 15-50/80-120, 3 mm slice thickness and 44 contiguous axial slices), and b) conventional SE T1-weighted (TR 600-650, TE 10-20, 3 mm slice thickness and 44 axial slices) five minutes after the injection of 0.1 mmol/kg Gd. Except for the screening scan and the week 16 or end-of-study scans, unsatisfactory images were rejected but not repeated because of the frequent scanning interval. Lesions were outlined and lesion volumes measured as described previously. Baseline normalized brain volumes (NBV) and percentage brain volume changes (PBVC) over the study period were assessed using SIENA and SIENAx software.

Statistical analysis

Due to the non-parametric nature of the statistical analysis of the primary target variable (cumulative number of newly active lesions after 16 weeks), the approach of Noether et al. was used for sample size estimation. 80% power for a Wilcoxon-Mann-Whitney test with 2  41 patients and one-sided alpha of 0.024 can be achieved when assuming that the probability - for a randomly chosen patient - of showing fewer cumulated newly active lesions under BX471 compared to placebo is 68%.1 This value is in line with the effect of IFN beta-1b in a placebo-controlled study.2 Since we were using the O'Brien–Fleming alpha spending function, the sample size was increased by multiplying with factor 1.008, ending up with 42 patients. Additionally, an overall drop-out rate of about 10% was expected and therefore a minimum of 47 patients per group, i.e. a total of 94 patients were to be randomized.

A pre-planned efficacy interim analysis was performed after 45 patients with a one-sided nominal interim alpha of 0.00127 and futility stopping if the one-sided p-value were ≥0.5. As the Independent Data Monitoring Committee did not recommend stopping the trial at the interim stage, the final analysis was performed with a one-sided nominal alpha of 0.02407. Analyses were performed on the “full analysis set” (all randomized patients) and on a subset called “valid case analysis set” which included patient data prior to either of the following events: 1) less than 50% of the scheduled drug being taken; 2) treatment with any pharmacologically “active” MS therapy other than steroids.

A central randomization list using permutated blocks of size 4 (balanced for treatment) was generated by the study’s Central Randomization Group. A centralized telefax was used to implement the random allocation. Neither patients, investigators nor the central MRI assessors were aware of the treatment allocation during the entire trial period.

A covariate-adjusted non-parametric analysis of covariance with baseline adjustment was used as the primary efficacy analysis of the cumulative number of newly active lesions at week 16. Patients who had undergone no baseline scan or fewer than 50% of the post-baseline scans were not considered for the primary efficacy analysis. Missing data rules based on the assumption of a linear increase of lesion counts over time were used to construct the cumulative number of newly active lesions at week 16 in patients who did not have all of the post-baseline evaluations.

Exploratory non-parametric analyses of covariance were also applied for other MRI parameters, and general linear models for repeated measures or pairwise comparisons for different time points on the basis of a priori hypotheses for immunological evaluations (SPSS®).

1. Noether GE. Sample-Size Determination for Some Common Nonparametric-Tests. Journal of the American Statistical Association 1987; 82:645-647.

2. Miller DH, Molyneux PD, Barker GJ et al. Effect of interferon-beta1b on magnetic resonance imaging outcomes in secondary progressive multiple sclerosis: results of a European multicenter, randomized, double-blind, placebo-controlled trial. European Study Group on Interferon-beta1b in secondary progressive multiple sclerosis. Ann Neurol 1999; 46:850-859.