Ocrelizumab versus Interferon Beta-1a in Relapsing Multiple Sclerosis
2017 · paper · Clinical trials
Two identical phase 3 trials comparing ocrelizumab with interferon beta-1a in 1,656 patients with relapsing multiple sclerosis over 96 weeks. Ocrelizumab lowered annual relapse rates by about 46–47% and reduced disability progression and MRI lesions.
3 findings
Substance Methods define NEDA as "the proportion of patients with a baseline EDSS score of at least 2.0 who had no evidence of disease activity", but Table 2 evaluates more patients for NEDA (382/384/379/375) than for disability improvement, which is restricted to baseline EDSS ≥2.0 (310/306/318/308).
- p. 3 Trial Procedures and End Points, 10th secondary end point
the proportion of pa-tients with a baseline EDSS score of at least 2.0 who had no evidence of disease activity
- p. 6 Table 2, NEDA 'No. of patients evaluated' vs disability improvement 'No. of patients evaluated'
382
Only the authors can say what is right here.
Judge's note: Page 3 shows the EDSS ≥2.0 qualifier on the NEDA end point. Table 2 (p6) restricts disability improvement to EDSS ≥2.0 (footnote †) and evaluates 310/306/318/308 patients there. NEDA counts are larger in every arm. The NEDA footnote ** (p7) defines the population as modified ITT and does not mention EDSS. The restriction stated in the Methods cannot have been applied to the NEDA counts shown.
Second check: another model agreed
Report a problem with this finding
Reference or label "B-cell reconstitution and preexisting humoral immunity should be relatively preserved with ocrelizumab treatment.26,27,34" cites an alemtuzumab trial (26), a Rebif neutralizing-antibody review (27) and the Rebif SmPC (34).
- p. 12 Discussion, paragraph on B cells
preserved with ocrelizumab treatment.26,27,34
- p. 14 References 26, 27, 34
Neutralizing antibod-ies in MS therapy: reviewing the Rebif experience
- p. 14 References 26, 27, 34
Alemtuzumab for patients with relapsing multiple sclerosis after disease-modifying therapy
- p. 2 Introduction, paragraph 2
stitution and preexisting humoral immunity.13,14
Only the authors can say what is right here.
Judge's note: Reference list on p14 checked: 26 Coles (alemtuzumab CARE-MS II), 27 Al-Sabbagh (Rebif neutralizing antibodies), 34 Rebif EPAR. None concerns CD20 or humoral immunity. The same claim in the Introduction (p2) cites refs 13 and 14 (B-cell immunobiology; plasma-cell maintenance during CD20 therapy, p13). Refs 26 and 27 are also cited in the previous paragraph for interferon neutralizing antibodies.
Second check: another model agreed
Report a problem with this finding
Reference or label Results say "In the intention-to-treat population in the OPERA I trial, 47.9% ... had no evidence of disease activity", but the Methods and Table 2 footnote ** say NEDA was "assessed in a modified intention-to-treat population" (382/384/379/375 evaluated vs ITT 410/411/417/418).
- p. 8 Results, Disability section, NEDA paragraph
In the intention-to-treat population in the OPERA I trial, 47.9% of the patients in the ocrelizumab group had no evidence of disease activity
- p. 3 Statistical Analysis
for the end point of no evidence of disease activity, in a modified inten-
- p. 7 Table 2 footnote **
the end point of no evidence of disease activity was assessed in a modified intention-to-treat population
- p. 8 Results, NEDA paragraph
In the intention-to-treat population in the
Judge's note: Pages 3–4 Statistical Analysis and the p7 footnote ** define the NEDA analysis set as modified ITT. Table 2 (p6) shows fewer patients evaluated than the ITT Ns. The p8 text labels the population as intention-to-treat, so the population label is wrong.
Second check: another model agreed
Report a problem with this finding
6 minor (typos and the like; most checked by the judge only)
Wording Methods define NEDA as including "no disability progression confirmed at 12 weeks or at 24 weeks"; Table 2 footnote ** says "no events of disability progression confirmed at 12 weeks".
- p. 3 Trial Procedures and End Points
no disability progression con-firmed at 12 weeks or at 24 weeks
- p. 7 Table 2 footnote **
no events of disability progression confirmed at 12 weeks
Judge's note: Both wordings are present on pages 3 and 7. Progression sustained for 24 weeks is also sustained for 12 weeks, so the two definitions are probably equivalent in practice. The wording still differs, so this is recorded as a minor consistency point.
Second check: another model agreed
Report a problem with this finding
Wording NEDA is the 10th of "10 hierarchically ordered secondary end points" in the Methods, but Table 2 lists it under "Exploratory end point" and the Results say "by 96 weeks (exploratory end point)".
- p. 3 Trial Procedures and End Points
There were 10 hierar-chically ordered secondary end points
- p. 6 Table 2 section header
Exploratory end point
- p. 8 Results
by 96 weeks (exploratory end point)
Only the authors can say what is right here.
Judge's note: Page 3 lists NEDA last in the 10 hierarchical secondary end points. Page 4 confirms 10 secondary end points. Page 8 also calls it one of the 'subsequent secondary efficacy end points', yet a few lines later labels it exploratory, as does Table 2 (p6). The same end point therefore carries two different tiers.
Second check: another model agreed
Report a problem with this finding
Wording Table 2 labels the MSFC values "Adjusted mean score at wk 96", while the abstract and Results describe the same values (0.21/0.28 vs 0.17) as the change from baseline to week 96.
- p. 6 Table 2, MSFC score row
Adjusted mean score at wk 96
- p. 8 Results text
The difference in the adjusted mean change in the Multiple Sclerosis Functional Composite score from baseline to week 96
Judge's note: Table 2 (p6) row label as quoted. The abstract (p1) says 'The change in the ... score' with 0.28 vs 0.17 and 0.21 vs 0.17. Page 8 says 'adjusted mean change'. The footnote § says scores were adjusted for baseline MSFC, so the between-group differences match under either reading. The row label still describes a different quantity from the text.
Second check: another model agreed
Report a problem with this finding
Typo Reprint address reads "Petersgraben 4, CH-4301 Basel, Switzerland"; the University Hospital Basel postal code is presumably CH-4031.
- p. 1 Correspondence address, side column
4, CH-4301 Basel, Switzerland
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the postal code of University Hospital Basel, Petersgraben 4, is 4031, not 4301. The crop shows 'CH-4301 Basel' as quoted. The page 13 affiliations give no postal code to compare against.
Second check: none (judge only)
Report a problem with this finding
Typo Discussion has "(F. Hoff-man–La Roche data on file)"; the sponsor is spelled "Hoffmann–La Roche" elsewhere.
- p. 12 Discussion, PML sentence
(F. Hoff-man–La Roche data on file)
- p. 2 Methods, Trial Oversight
The sponsor, F. Hoffmann–La Roche, designed
- p. 12 Discussion, PML sentence
F. Hoff-man–La Roche data on file
Judge's note: The p12 image shows 'Hoff-/man–La Roche' with a single n. Pages 1, 2 and 13 spell it 'Hoffmann'.
Second check: none (judge only)
Report a problem with this finding
Typo Reference 27 prints "Al-SabbaghA." with no space between surname and initial.
- p. 14 Reference 27
Al-SabbaghA.
- p. 14 Reference 27
Al-SabbaghA. Neutralizing antibod-
Judge's note: The p14 image shows 'Al-SabbaghA.' run together. Other references separate surname and initials.
Second check: none (judge only)