Intended for healthcare professionals

Rapid response to:

Analysis And Comment Drugs

Lessons for clinical trials from natalizumab in multiple sclerosis

BMJ 2006; 332 doi: (Published 16 February 2006) Cite this as: BMJ 2006;332:416

Rapid Response:

Author's reply

I thank the authors of rapid responses for their interest shown in my
article. It is difficult to offer a brief response to the issues raised.
In the AFFIRM trial(1), relapse of multiple sclerosis was observed in 6%
of patients in the natalizumab group and 13% of placebo group;
corresponding figures in the SENTINEL trial for natalizumab plus
interferon beta-1a and interferon beta-1a alone were 5 and 9%
respectively(2). This would suggest that between 15 to 25 patients with
relapsing multiple sclerosis would require to be treated for at least two
years with natalizumab to prevent one relapse. The cost of the drug itself
would be around US $ 1 million to prevent one relapse with an estimated
number-needed-to-treat (NNT) of 20.

The rebound effect of increased relapse rates after discontinuation
of natalizumab (but not of placebo) is also of concern. In the AFFIRM
trial the annualised relapse rate rose from 0.26 relapses to 0.495 for
patients receiving natalizumab, but the annualised relapse rate dropped
from 0.81 to 0.608 in the placebo group (1). The data on relapse rates
after discontinuation of natalizumab were not revealed in the SENTINEL
trial (2). The lack of sustained clinical effect of natalizumab casts
doubt about its ability to modify actual disability progression. The two-
year disability progression in natalizumab trials was based on complex
statistical modelling, rather than observed differences in the Expanded
Disability Status Scale (EDSS) scores of patients compared between two
time points (entry and end of trial). For reasons unknown, the EDSS scores
of patients were not revealed at the end of study period. The probability
of 54% risk reduction in disability progression of patients on natalizumab
in the AFFIRM trial appears to be a highly skewed statistical projection
given a relative difference of 7% in the relapse rate between the active
and the placebo arms (1). The statistical analysis was performed by the
sponsors (Biogern Idec and Elan Pharmcaeuticals) and not by an independent
group. The estimated disability progression in these trials is not a true
measure, but a probability that remains to be validated in the
longitudinal history of relapsing-remitting multiple sclerosis.

The responses from Giovannoni and others, and also from Hutchison and
Tubridy, question my comparison of the placebo arm of natalizumab
monotherapy in the AFFIRM trial (1) and beta-interferon-1a arm of the
SENTINEL trial. Patients in the SENTINEL trial (2) had relapsing-remitting
multiple sclerosis, not relapsing-progressive multiple sclerosis. The most
significant differences between the two groups was in the MRI appearances
of enhancing lesions (which are known to be influenced by pre-treatment
with beta-interferon) and mean age group of recruited patients but
otherwise base line characteristics were comparable for most clinical and
demographic data, such as age range, sex, race and number of relapses in
the previous year. Baseline mean EDSS scores of the interferon beta-1a
alone arm in the SENTINEL trial (2) and placebo group of the AFFIRM trial
(1) were 2.5 (range 0-5.5) and 2.31 (range 0-6) respectively. My view
regarding the placebo-like effect of beta-interferon 1a in the SENTINEL
trial has also been echoed in the editorial commentary on these two trials

The risk of progressive multifocal leukoencephalopathy (PML) as a
direct adverse effect of natalizumab therapy is uncertain but is likely to
be significant (4) and higher than 1 in 1000 if patients receive treatment
beyond 18 months. The major limitation of the safety evaluation of
natalizumab, as recognised by the authors of this study (5), was that the
safety data were collected only after discontinuation of natalizumab and
not when patients were still receiving therapy. While the safety
evaluation was possible in 97% of patients with multiple sclerosis within
three months after the suspension of natalizumab in 28 February 2005, it
is not clear how many patients were actually screened within three months
of their last natalizumab infusion.

Natural history of multiple sclerosis suggests that for most patients
it is a relatively slowly progressive disease (6). The key point of my
article was that short-term trial results could not be extrapolated to
long-term safety and effectiveness with specific reference to natalizumab
in multiple sclerosis. Previous studies of natalizumab did not confirm its
clinical effectiveness. MRI based outcome measures (gadolinium enhancing
lesions) and 6-month relapse rates are not predictive markers for long
term disability in multiple sclerosis. Fast-track approval of natalizumab
by the FDA before the publication of the AFFIRM and SENTINEL trials did
assume its relative safety. Subsequent events have proven this assumption
to be incorrect. In my view, the unknown risk of PML, development of
neutralising antibody limiting therapeutic effect (6-12% of all patients
after two years), adverse events such as fatigue and infection, a possible
increase of breast cancer cases (none in placebo group, three cases in the
treatment group in the AFFIRM trial) and rather dubious statistical data
on disability progression are not sufficiently persuasive to recommend
natalizumab as a safe and an effective therapy for general patients with
multiple sclerosis, especially in the absence of a comparative study with
pulse methylprednisolone.

While I appreciate that the responses from the investigators of
AFFIRM trial reflect purely research and clinical interest, it may invite
an uncomfortable but inevitable question whether one of the aims of
pharmaceutical sponsorship (conflict of interest statement, AFFIRM trial1)
is to create academic guardians to protect interest of the industry.


1. Polman CH, O¡¯Connor PW, Havrodva E, et al. A randomised, placebo-
controlled trial of natalizumab for relapsing multiple sclerosis. N Engl J
Med 2006; 354:899-910.

2. Rudick R, Stuart WH, Cababresi PA, et al. Natalizumab plus
interferon beta-1a for relapsing multiple sclerosis. N Engl J Med 2006;
354: 911-23.

3. Ropper AH. Selective treatment of multiple sclerosis. N Engl J Med
2006; 965-7.

4. Langer-Gould A, Steinman L. What went wrong in the natalizumab
trials? Lancet 2006; 367: 708-10.

5. Youary TA, Major EO, Ryschkewitsch C, et al. Evaluation of
patients treated with natalizumab for progressive multifocal
leukoencephalopathy. N Engl J Med 2006; 354: 924-33.

6. Benedikz J, Stefansson M, Guomundsson J, et al. The natural
history of untreated multiple sclerosis in Iceland. A total population-
based 50 year prospective study. Clin Neurol Neurosurg 2002; 104: 208-10.

Competing interests:
None declared

Competing interests: No competing interests

15 March 2006
Abhijit Chaudhuri
Consultant Neurologist
Essex Centre for Neurological Sciences