Neurology -- Correspondence for CHAMPIONS Study Group, 66 (5) 678-684

Correspondence published:

IM interferon ß-1a delays definite multiple sclerosis 5 years after a first demyelinating event: Douglas S Goodin (5 June 2006)

Reply from the authors: Revere P Kinkel, Craig Kollman (5 June 2006)

IM interferon ß-1a delays definite multiple sclerosis 5 years after a first demyelinating event 5 June 2006

Douglas S Goodin,
University of California, San Francisco
505 Parnassus Ave, Rm M794, San Francisco, CA 94143-0114
Send Correspondence to journal:
Re: IM interferon ß-1a delays definite multiple sclerosis 5 years after a first demyelinating event

douglas.goodin{at}ucsf.edu Douglas S Goodin

The recently published CHAMPIONS trial [1], reported that the early use of weekly IFN beta-1a (compared to delayed treatment) reduced the likelihood of developing clinically definite (CD) multiple sclerosis (MS) after a 5-year follow-up period in patients who had initially presented with clinically isolated syndromes (CIS) suggestive of MS. There are concerns both about the trial design and this conclusion.
For example, this trial was planned only after the completion of the CHAMPS trial [2] and with full knowledge of the CHAMPS results. Additionally, the primary endpoint of the CHAMPIONS trial was the development of CDMS at any time after CIS-onset. Consequently, at the outset of CHAMPIONS, there was already a highly significant difference (bias) between groups with respect to the primary outcome measure. Significantly more patients in the ‘delayed-treatment’ group had already reached (and were known to have reached) their final endpoint before the trial even began compared to the ‘early-treatment’ group. Conversely, there were significantly more ‘survivors’ (i.e., those without a second clinical attack) in the ‘early-treatment’ group.
Therefore, as a consequence of this experimental design, the final result of CHAMPIONS study was already assured before the trial even began, unless for some reason the rate of development of CDMS happened to be significantly greater in the ‘surviving’ patients of the ‘early-treatment’ group compared to the ‘delayed-treatment group’.
During the CHAMPIONS trial, there was no suggestion of a difference between groups in the rate of development of CDMS (Figure 2 of the paper). Indeed, the principal study result can be completely attributed to the experimental bias and is not due not to any observations made during CHAMPIONS trial itself. This result should have been anticipated.
First, it is unclear how such an increased rate of CDMS in the ‘early-treatment’ group could have been explained, let alone anticipated. Second, if the known experimental bias is removed from the CHAMPIONS study by excluding those patients who had already reached their endpoint at study entry, the CHAMPIONS trial effectively boils down to an investigation of the likelihood of developing CDMS in two groups of patients, currently receiving active therapy, who have not had a second clinical attack 2-3 years after their initial clinical episode suggestive of MS. It seems hardly surprising that, at this point, early- versus delayed-treatment makes little or no difference.
References
1. CHAMPIONS Study Group. IM interferon beta-1a delays definite multiple sclerosis 5 years after a first demyelinating event. Neurology 2006;66:678-684.
2. Jacobs LD, Beck RW, Simon JH, et al. and the CHAMPS Study Group. Intramuscular interferon beta-1a therapy initiated during a first demyelinating event in multiple sclerosis. N Engl J Med 2000;343:898- 904.
Disclosure: The author reports no conflicts of interest.

Reply from the authors 5 June 2006

Revere P Kinkel,
Beth Israel Deaconess Medical Center
330 Brookline Ave, Boston MA 02215,
Craig Kollman
Send Correspondence to journal:
Re: Reply from the authors

rkinkel{at}bidmc.harvard.edu Revere P Kinkel, et al.

We thank Dr. Goodin for his comments on our article and the opportunity to further discuss this important issue. He is correct that the CHAMPIONS study was planned with full knowledge of the CHAMPS outcome, but the primary outcome in the CHAMPIONS study was determined by the same group of blinded, independent outcome committee members that determined this outcome in the original CHAMPS study.
We disagree with Dr Goodin’s assertion that the differences between treatment groups at CHAMPIONS onset (i.e. at the end of CHAMPS) assured the study outcome. Although the CHAMPS study reported that AVONEX reduces the incidence of CDMS 2 years following CIS onset, this does not guarantee that the benefit of immediate treatment will last indefinitely, particularly when compared to an active treatment group. It is possible that the effect of AVONEX is only to delay the conversion to CDMS. If so, the incidence of CDMS in the immediate treatment group (originally randomized to AVONEX) will eventually “catch up” to that in the delayed treatment group (originally randomized to placebo). The primary question in the CHAMPIONS study was whether the benefit of immediate treatment would persist beyond 2 years.
Dr Goodin suggests that there is no reason to expect the ‘surviving’ immediate treatment patients at the end of CHAMPS to develop CDMS during CHAMPIONS at a rate greater than the ‘surviving’ delayed treatment patients. In fact this is exactly what one would expect if the higher risk immediate treatment patients ‘survived’ at a greater rate until the end of CHAMPS because the treatment was effective. For the same reason, we disagree with his interpretation of Figure 2. Kaplan-Meier curves that eventually run parallel to one another do not imply a loss of treatment effect.
Dr Goodin suggests that a more appropriate analysis would be to exclude patients who already developed CDMS during CHAMPS, essentially looking at the conditional probability of CDMS at 5 years given that it had not yet occurred at 2 years. This analysis would be flawed on both clinical and statistical grounds. Comparison of conditional rates would be of little value clinically since the decision whether to treat immediately with AVONEX must be made at CIS onset, not 2 years later.
Furthermore, excluding patients who converted to CDMS during CHAMPS would selectively remove more of the higher risk cases from the denominator of the original placebo group leaving inherently different cohorts of patients. This would confound any comparison of subsequent failure rates.
Disclosure: The study mentioned in this Correspondence was supported by Biogen Idec, Inc. Drs. Kinkel, Kollman, O’Connor, Murray, and Simon have received grants and honoraria from Biogen Idec, Inc.