Analysis
of the results of a randomized study of hyperbaric oxygen
therapy in
the treatment of children with cerebral palsy:
Placebo or
physiological effect?
Pierre
Marois, MD,
FRCP(C), Physiatrist and Michel Vanasse, MD, FRCP(C), Neurologist
Introduction
The results
of the study of the effects of hyperbaric oxygen therapy (HBO study) in the
treatment of children
with cerebral
palsy have been known for over six months. However, these results are still
controversial. The
directors of
the "Fonds de la recherche en santé du Québec" (FRSQ), the
organization responsible for
overseeing
this study and some researchers, including the principal researcher, Dr.
Jean-Paul Collet, are
convinced
that the improvements observed in both groups of children involved in the study
was due solely to a
placebo
effect. Others, the authors included, believe that the cause or causes of this
improvement were not
clearly
identified by this study and suggest the need to pursue further research of HBO
treatment.
We have
voiced our opinions on several occasions and are now compelled to document our
position in order to
enable those
interested in this debate to familiarize themselves with our argument, weighing
the pros and the
cons. We
would like to demonstrate that we are not merely presenting an empathetic
attitude towards desperate
parents faced
with the unfortunate condition of their child. Recognizing our limitations in
the area of research,
we associated
ourselves with and collaborated with a team of researchers, clinicians and
methodologists
possessing
the competence required to write the original version of this study, which was
left practically
unchanged by
Dr. Collet. However without prejudice, we feel that our twenty years of
experience in paediatric
rehabilitation
and our involvement in numerous research projects often dealing with
experimental treatments
gives us the
right to voice our opinion. We invite the readers to judge these objective
arguments on their own
merit.
A.HBO
Study: The uncontested facts
The consensus
of the research group was that the randomized HBO study for children with
cerebral palsy
demonstrated
that:
-Both groups
of children that participated in the study showed improvements in mobility as
measured by the
Gross Motor
Function Measure (GMFM), in language skills and in auditory and visual memory.
-The results
did not show that the treatment at 1.75 ATA and 100% O² were superior to the
so-called "placebo"
at 1.3
atmospheres with ambient air. As often pointed out while the research protocol
was being developed and
in the
discussion of the results, the "placebo" was in fact not an inert
treatment. We had recommended a true
control group
that would not have received any treatment.
Furthermore:
-The research
group and the scientific advisory committee were of the unanimous opinion that
the
improvements
observed were clinically significant.
Page 1 of 7
Result_En
6/12/01
http://www.marieenfant.qc.ca/Result_En.html
-The
improvements in the GMFM scores in both groups were observed during the retest
three months posttreatment.
-There were
no serious side effects in the course of the 2,500 HBO treatments administered
during the pilotstudy
and the
double blind study.
-All who
evaluated this study acknowledged its scientific merit.
B. The
reasons for the improvement
During three
long plenary sessions and several other meetings in small groups, the
researchers debated the
possible
explanations for the improvements observed in both groups of children We feel
it is appropriate to
convey the
researchers conclusions with respect to the possible reasons for the observed
improvements.
1. A placebo
and/or participation effect. According to many researchers the apparent cause
of observed
improvements
is the fact that the individual participated in a study. Without being able to
clearly identify the
mechanism,
studies have shown that the involvement in a research study resulted in
improvements in the
parameters
being evaluated. To explain the observed changes, we have made reference to the
placebo effect,
the Hawthorne
effect, the increased amount of time the parents spent with their children or
the increase in the
social
interaction of the parents and children.
2. The effect
of the HBO therapy. As mentioned above, the placebo referred to in this study
was not a true
placebo. It
was a HBO treatment at 1.3 atmospheres of pressure with ambient air. This
treatment has an effect
on the
partial pressure of blood gases and perhaps other physiological effects. It is
therefore possible that this
treatment
could have a beneficial effect at the cerebral level.
Hypotheses
not retained by the research team
The team of
researchers evaluated the possibility that the improvements were due to the
natural evolution in
the
neurological condition of the subjects or the effect of learning in the tests
used. These hypotheses were not
retained as
the children’s development was stable or had shown only slight improvement. The
improvement
was much
greater following the HBO treatment. We also consider that there was no
learning effect in the tests
used,
especially with the GMFM, which was the principal variable.
C. Our
Position
Our opinion
is based on the following points that we will elaborate:
-The placebo
was not in fact inert and it would be more appropriate to claim that these
children received a
reduced dose
of HBO, in other words an exposure to a pressure of 1.3 atmospheres of ambient
air.
-Without
denying the existence of a placebo effect, or the possibility that the observed
improvements are due
to the
placebo effect, we maintain that there has been no scientific proof that:
-a placebo
can significantly improve mobility, language and memory of children with
cerebral palsy clinically
or
statistically. If the results were due to a placebo, then it was as effective
on motor performance as the
conventional
therapy currently considered the most effective treatment (intensive physical
therapy) and in a
shorter
amount of time.
-a placebo
can produce long-lasting effects. The improvements were still observed three
months post-treatment
in the retest
of the subjects involved in this study. Without being able to scientifically
confirm, the impression
of both
parents and health care professionals following these children is that the
improvements persisted, (one
year and
more).
Page 2 of 7
Result_En
6/12/01
http://www.marieenfant.qc.ca/Result_En.html
-a placebo
can result in reproducible improvements from one study to another. The mobility
improvements
measured by
the GMFM in both groups involved in the randomized study were identical to the
results of three
pilot studies
and one case study. We are in agreement with the experts on the scientific
advisory committee
that there is
no irrefutable scientific proof demonstrating any therapeutic benefit of
low-pressure hyperbaric
therapy. This
however does not imply that a beneficial effect does not exist. Some clinical
and experimental
data allows
us to defend this hypothesis.
D. The
"placebo" group received a reduced hyperbaric treatment
That the
placebo was not inert was not only a concern upon completion of the study, but
had been pointed out
and discussed
during the development of the research protocol. The initial versions of the
protocol (versions
that were
submitted and approved by the ethical committees of Hopital St. Justine, Centre
hospitalier regional
de Rimouski
and Institut en réadaptation en déficience physique de Québec in May 1999) read
as follows:
"Since a
pressure of 1.3 atmospheres could eventually produce an effect (it is not an
inert placebo), we had
proposed the
introduction of a third group in the study that would not receive any
treatment. The children in
the control
group would be evaluated, with the awareness that they didn’t receive the
treatment, initially, one
month later
and at the end of the study. This group of subjects would allow us to verify
the natural evolution of
the condition
and the effect of repeating the evaluation on the measured performance."
On Dr.
Collet’s recommendation (eventually approved by the majority of researchers)
the inclusion of this
third control
group was discarded in order to include more subjects in the other two groups
thus obtaining more
significant
statistics.
We were not
the only ones to have raised this methodological problem. The results of this
study were
presented at
the Underwater and Hyperbaric Medical Society Conference by
Dr. Stéphane
Tremblay in June 2000. The report of Dr. Tremblay’s presentation states:
"many individuals
recommend verifying
in future studies whether in fact the placebo was inert (28% O2=placebo) as
well as the
effect of
oxygen at partial pressure, seeing as the improvement in both groups was
clinically significant." This
point was
also raised during the presentation of the results of this study at the
American Academy of Cerebral
Palsy and
Developmental Medicine in Toronto, September 2000, given in part by Dr. Michel
Vanasse.
A letter
addressed to Dr. Collet from Dr. Butcher, the senior editor of the journal The
Lancet, accepting the
text of the
results of our study stated:
"I am
very pleased to be able to tell you that The Lancet’s editorial team had
decided to accept your paper for
publication
provided that all references to "placebo" are changed to
"slightly pressurized room air (or
something
similar)"."
Evidently
these comments do not allow us to conclude that the treatment was responsible
for improvements
observed in
the group that received the attenuated hyperbaric therapy. However, it does
indicate that many
scientists
feel this possibility cannot be excluded and should be the focus of further
investigation. It was the
conclusion in
the resume of the paper written by Dr. Collet, on the results of our study that
has been accepted
for
publication in The Lancet:
"The important
improvement observed in both groups for all three dimensions tested deserves
further
considerations."
E. Comments
regarding the "placebo" effect or the effect of participation
1. Placebo
effect, effect of participation (inclusion benefit) and "Hawthorne
effect"
Page 3 of 7
Result_En
6/12/01
http://www.marieenfant.qc.ca/Result_En.html
Without
intending to do a complete analysis of the placebo effect, we feel it is
important to comment on the
subject.
During the discussions pertaining to the possible causes of the improvements
noted in both groups of
subjects,
many of the researchers
maintained
that it was due to a placebo effect, inclusion benefit or "Hawthorne
effect" or a combination of the
three. The
designated committee of experts chosen by the FRSQ concluded that this was the
most probable
explanation
of the reported improvements. Some reproached us for attempting to deny the
obvious, that these
effects are
well known and well documented. If this is in fact the case, it is surprising
to report that the paper
reporting the
results of our study only includes one very general reference to this subject.
We do not in any
way deny the
existence of the placebo effect, however, we have always maintained that there
is no scientific
proof that
such an effect could explain the improvements reported in the children that
received either the
complete or
reduced hyperbaric therapy. The placebo effect is a well-known phenomenon to
health care
professionals
that has rarely been studied. The titles of certain papers cited by the experts
of the scientific
committee
refer to these studies: "The Mysterious Placebo Effect", "The
Uncontrollable Placebo Effect".
The effect of
participating in a study as possibly being the sole factor resulting in an
improvement in the
subjects
involved was also postulated. It is a fascinating hypothesis and remains just
that, a hypothesis. The
effect of
participating in a study, also referred to as the inclusion benefit, is not
well known. Remarks made by
Dr. Lantos
suggesting this phenomenon remains hypothetical in an editorial in the Journal
of Pediatrics 1999,
"The
"inclusion benefit" in clinical trials":
"As a
thought experiment, let us suppose that it is really true that participants in
randomized clinical trials do
have better
outcomes than similar patients with similar diseases treated in the same
institution at the same
time".
In the same editorial he adds: "The phenomenon of inclusion benefit, if
real".
The
"Hawthorne " effect was also proposed as a possible cause for the
reported improvements of the subjects
in our study.
This effect is defined as: the effect of being conscious of being observed can
cause a modification
in the
behavior of an individual. In a recent study of the "Hawthorne"
effect and the sensation of feeling better
after
anesthesia, De Amici et al. (2000) concluded: "Whereas this study answers
the question concerning the
importance of
the Hawthorne effect in a field where subjective perception is predominant, the
impact of this
phenomenon on
more "objective " parameters remains open. However, the improvements
reported in the
children
treated were based on objective data, not on a sensation of feeling better.
These remarks can explain
our
reluctance to believe that the placebo, the participation or Hawthorne effects
be the cause of the reported
improvements.
Furthermore:
2. A placebo
as effective or even more so than conventional treatment?
We have often
mentioned that there are few effective treatments for children with cerebral
palsy, and thus the
importance of
not neglecting a treatment as promising as hyperbaric therapy. In the past few
years there have
been several
papers investigating therapeutic modalities available for children with
cerebral palsy, which have
concluded
that there is no scientific evidence to support the effectiveness of these
treatments. (Turnbull 1993,
Graves 1995,
Majnemer 1998). Dr. Majnemer wrote in her paper: "there is a lack of evidence
to support the
efficacy of
rehabilitation interventions in children with cerebral palsy".
Some open or
pilot studies have reported that intensive physical therapy (6 to 8 months, two
times a week)
result in a
functional improvement measured by the GMFM (Gross Motor Function Measure, the
evaluation
tool used in
our pilot and randomized studies). However, for comparable populations,
children with spastic
diplegia, we
observed a 5.3% improvement on the global score of the GMFM in our pilot study
(after 20
treatments,
after one month) compared to an improvement of 4% reported by Russell et al. (
after 6 months of
intensive
therapy). Furthermore, in the subjects involved in the randomized study, a
generalized improvement
was reported,
including attention and communication.
Page 4 of 7
Result_En
6/12/01
http://www.marieenfant.qc.ca/Result_En.html
We can
therefore establish that the hyperbaric therapy resulted in functional
improvements more rapidly and
more
generalized than conventional treatment. If we accept that the improvement
observed in the children
having
received HBO therapy is due to a placebo, must we then conclude the
improvements resulting from the
6-8 months of
intensive physical therapy were also due to placebo because the results were
identical?
3.A
reproducible and persisting placebo effect?
Another
interesting and, in our opinion, very important element that was highlighted by
our research was that
the
improvements persisted at least three months post treatment. The children were
systematically re-evaluated
three months
later and we were able to document beyond doubt the persistence of the gains
observed after 40
hyperbaric
treatments (and even a slight improvement which was not statistically
significant). In a pilot study
Dr. Maurine
Packard of Cornell University evaluated 26 children that had received 40 HBO
treatments (each
lasting one
hour at 1.5 ATA). She noted that the improvements in the areas of attention,
language and ability to
play that
were observed immediately after the treatments were still present six months
later.
Our clinical
experience leads us to believe that the improvements reported persist even in
the long term, in
other words
for at least one year after the treatment. To our knowledge, no scientific
proof exists confirming
the
persistence of a placebo for that period of time.
We are aware
that there does not exist, other than our research study, a randomized, double
blind study
conducted to
analyze the effect of this therapy. However, there have been some credible case
studies and pilot
studies on
children with cerebral palsy (Paleg 1998, Barrett 1999, Montgomery et al. 1999,
Packard 2000),
published or
presented at conferences. We may have reservations about the conclusions of
these studies, but
one thing
does remain unchanged, they all showed similar improvements.
F. Is a
low-pressure hyperbaric treatment effective?
We would like
to restate that we are in agreement with the report of scientific advisory
committee that there is
no
irrefutable proof illustrating that a HBO treatment at low pressure is
effective on humans. Equally, there is
no scientific
proof showing that it is ineffective as there has never been any research on
humans to evaluate
this
hypothesis. In our opinion, some of the clinical and experimental data deserve
consideration.
-a recent
study evaluating hyperbaric treatment with and without oxygen therapy in the
treatment of cerebral
vascular
accidents in rats showed that: " …Hyperbaric oxygen and, to some extent,
hyperbaric pressure
reduced
ischemic brain damage and behavioral dysfunction."(Chang et al. 2000)
-A
low-pressure hyperbaric treatment ( without oxygen) , the Gammow bag, is
effectively used to treat cerebral
edema
suffered by some individuals when at high altitude, a condition known as
"acute mountain sickness".
(Austin 1998)
It is estimated that the pressure reached with this bag is approximately .2
ATA. Recently,
Heuser et al.
reported a clinical improvement after ten hyperbaric treatments at a pressure
of 1.3 ATA and in
the cerebral
SPECT scan in six patients presenting a toxic encephalopathy.
-Two
double-blind studies were conducted on humans to evaluate the effectiveness of
HBO therapy in the
treatment of
CVA in the acute phase. (Anderson et al. 1991, Nighoghossian and Trouillas
1995) The scientific
advisory
committee did not cite these studies. In both studies, the experimental group
received a treatment of
hyperbaric
oxygen therapy at 1,5 ATA with 100% oxygen and the placebo group 1.5 ATA
without oxygen.
The study by
Nighoghossian and Trouillas reported an improvement in both groups of subjects,
with a greater
improvement
(however not statistically significant) in the group of patients that received
the hyperbaric
treatment
with oxygen. In the study conducted by Anderson et al. improvement was reported
in both groups,
however the
greater improvement was noted in the group that received hyperbaric therapy
without oxygen
(the
difference was not significant statistically between the two groups). These
patients were evaluated with a
quantified
neurological evaluation and a measurement of the volume of the cerebral infarct
using repeated
cerebral
scans. The patients in the "placebo" group experienced more favorable
results in these two
Page 5 of 7
Result_En
6/12/01
http://www.marieenfant.qc.ca/Result_En.html
measurements,
suggesting that improbability of a placebo effect.
It is not up
to us to evaluate in hindsight the results of these two studies. However, it
would have been very
interesting
to compare the evolution of both groups with a control group without treatment.
In our opinion, the
results of
both of these studies, especially the one by Anderson et al., support the
possibility that a hyperbaric
treatment at
low pressure (1.5 ATA in that study) can have a therapeutic effect.
Furthermore, the physiological
effects of
hyperbaric treatment are certainly more complex than solely the increase in
partial pressure of
oxygen (Buras
2000). It would be interesting to study the physiology of hyperbaric therapy
with and without
oxygen
therapy.
G. Conclusion
The
conclusion and the subsequent proposals are taken verbatim from what was
proposed in May 2000. Our
position
remains unchanged with regard to the follow-up of our research project.
We are
convinced that our work on this therapeutic approach must be pursued in order
to identify the cause or
causes of the
reported improvements. It would be difficult to accept shelving the results of
this study. We base
our
convictions on the fact that the improvement in mobility reported in these
children after two months of
hyperbaric
treatment is similar to that measured after six months of intensive physical
therapy (twice a week).
Furthermore
the children involved in this study showed improvements in language and
neuropsychological
evaluations
as well as mobility. Finally, the improvements persisted for at least three
months post-treatment (as
verified
scientifically).
It is our opinion
that in order to properly investigate the different hypotheses raised, a
double-blind study must
be conducted
where a true placebo would be compared to different treatments (oxygen alone,
hyperbaric
treatment
alone and hyperbaric oxygen). It is clear that we could not recruit enough
subjects for such a
thorough
study in Quebec. A multi-center Canadian or international study would have to
be considered . This
study would
require many months or years to plan, finance and realize. Furthermore it would
require
hyperbaric
chambers that are not available at the present time in the Quebec’s public
health care system.
Finally, even
though we did not definitively identify the cause or causes of the
improvements, we are
convinced
that these results will encourage parents to continue to pursue hyperbaric
treatments for their
children. For
these reasons, we recommend that at least one or more hyperbaric treatment
centers remain open
in Quebec.
H. Proposal
to maintain one or more hyperbaric treatment centers in operation
For the above
mentioned reasons we propose:
-to maintain
one or more hyperbaric treatment centers in operation
-to make
these centers readily available to treat patients presenting a chronic,
non-progressive encephalopathy ,
at their own
expense
These centers
should assume, in part, the responsibility of systematically evaluating the
patients undergoing
hyperbaric
treatment with a validated detailed questionnaire such as the PEDI. Depending
on their means, they
should also
evaluate mobility, language and cognition with as many subjects as possible.
These centers should
also agree to
participate with any efforts to conduct a multi-centre Canadian or
international study.
If one or
more centers continue to operate and administer treatments in Quebec it would:
-avoid the
need for parents to travel to Ontario or elsewhere (with all of the expenses
and hardship it entails) in
order to
receive treatments often administered without any medical supervision nor pre
and post treatment
evaluation.
Page 6 of 7
Result_En
6/12/01
http://www.marieenfant.qc.ca/Result_En.html
-allow us to
continue to study the effects of this treatment. Although we are well aware
that it would not be a
double-blind
study, it seems clear that the systematic evaluation process that we recommend
is better than no
evaluation at
all.
-maintain the
centers best prepared to participate in a multi-centre study due to their
geographic location and
their ability
to treat many subjects at once.
Pierre
Marois, physiatrist
Michel
Vanasse, neurologist
Page 7 of 7
Result_En
6/12/01
http://www.marieenfant.qc.ca/Result_En.html