Intended for healthcare professionals

Rapid response to:

Education And Debate

Where is the evidence that animal research benefits humans?

BMJ 2004; 328 doi: https://doi.org/10.1136/bmj.328.7438.514 (Published 26 February 2004) Cite this as: BMJ 2004;328:514

Rapid Response:

There remains an urgent need for systematically reviewing the results of animal experimentation

Before conducting any kind of new research the results from existing
research should be taken into account. Modern research synthesis demands
that appropriate steps are taken to avoid bias and random error and that
the validity of the extant research base is critically examined;
systematic reviews offer a means of doing this [1]. Few argue with the
need to conduct systematic reviews of clinical research, and we suggest
the imperative for systematically reviewing the results of animal
experimentation is exactly the same [2]. That the UK Medical Research
Council asks for evidence of systematic reviews in clinical trials, but
not for animal experiments, suggests a double standard operates.

Drs Blakemore and Peatfield suggest that ethical committees are
responsible for establishing whether or not there is enough evidence to
proceed from animal research to clinical trials, while Dr Bakhle suggests
that grant reviewers undertake this task. There is no evidence that either
group currently does this, nor is it a feasible expectation. For example,
a recent systematic review of randomised controlled trials of fluid
resuscitation in animal models of uncontrolled bleeding identified 44
trials from as far afield as Israel, USA, Norway and Greece [3]. It is
very unlikely that ethics committees or reviewers could be expected to be
aware of this amount of information and it can take several months to
conduct a systematic review. This amount of time is simply unavailable to
grant reviewers and ethical committee members. Investigators applying to
do research should assess the preceding relevant research but too often
such assessments are haphazard, biased and incomplete. Ethics committees
and the like should be able to refer to systematic reviews of animal
studies but at present there are almost none to draw upon.

Some letters drew attention to selected areas where the animal and
human studies appeared to agree. Evidence that some animal work accurately
predicts the human studies does not detract from the need for systematic
reviews. In addition to casting light on the issue of generaliseability,
systematic reviews provide a means of finding out if the animal work was
concluded in a timely way or whether unnecessarily duplicative experiments
were conducted, whether the animal work correctly preceded the human
trials and so on. Systematic reviews may demonstrate heterogeneous
effects: these may be attributable to study design, between species
differences, or more subtle sources of variability. Investigation of the
causes of heterogeneity is an important aspect of the learning
opportunities provided by systematic review.

The six systematic reviews that we considered were not specially
selected; they were the only cases, identified through an objective search
strategy, in which systematic reviews of both the animal and human studies
were available for the same condition. Dr Macleod suggests that our study
is biased because the systematic reviews of the animal studies were only
conducted because the corresponding clinical trials failed. However, such
a bias could only occur if systematic reviews are done after the human
trials. No bias is possible if systematic reviews of animal studies
precede human trials, as we recommend.

Dr Coulter uses the example of use of antenatal steroids for preterm
labour as evidence that animal research benefits humans. We believe it
provides evidence of the need to conduct systematic reviews. A cumulative
meta analysis of the 14 human trials [4] indicates that from the first
successful trial by Liggins in 1972 there were over twenty years of
trials, all adding evidence for a successful therapy. Had the cumulative
meta-analysis been done prospectively, perhaps half of the human antenatal
steroid trials would have been avoided, as the evidence for benefit was
already over-whelming. We do not know whether the accumulating animal
experience for antenatal steroids would show the same pattern of growing
evidence, whether unnecessary and repetitive experiments were done or
whether animal experiments continued after the human evidence was already
established and, if they did, what those studies added to knowledge. The
call we make is for ongoing systematic reviews of a complete body of
animal research, with updated meta-analysis as appropriate, to ensure that
animal research is always up to date, not unnecessarily duplicative, and
always relevant.

References

1. Egger M, Davey Smith G, Altman DG. (Eds) Systematic reviews in
health care: meta-analysis in context. 2001, BMJ Books, London.

2. Pound P, Ebrahim S, Sandercock P, Bracken M, Roberts I. Where is
the evidence that animal research benefits humans? British Medical Journal
2004; 328: 514-7.

3. Mapstone J, Roberts I. Fluid resuscitation in haemorrhagic
hypovolaemia: evidence from animal models. J Trauma (in press).

4. Sinclair JC. Meta-analysis of randomized controlled trials of
antenatal corticosteroid for the prevention of respiratory distress
syndrome: discussion. Am J Obstet Gynecol. 1995; 173(1):335-44.

Competing interests:
As outlined on original paper

Competing interests: No competing interests

23 March 2004
Pandora Pound
Research Fellow
Shah Ebrahim, Peter Sandercock, Michael B Bracken, Ian Roberts
Department of Social Medicine, Bristol University, BS8 2PR