By Dominic Wilkinson, @Neonatalethics
Earlier this year, the Lancet published a trial (the ‘ACT’ trial) involving 100,000 babies at risk of being born prematurely in developing countries. Half of the mothers in the ACT trial did not receive a simple cheap medicine that has been previously shown in multiple trials and meta-analysis to reduce the risk of death for premature babies. From the ACT trial results, it appears that 89 additional babies died as a result of their mothers taking part in the trial.
Surely this is an egregious example of unethical research? It appears to be in breach of the World Medical Association Declaration of Helsinki standards. Why did ethics committees allow the research? Why did a major journal like the Lancet publish it? Why aren’t bioethicists and activist and advocacy groups like Public Citizen jumping up and down in outrage?
The Declaration of Helsinki appears to be clear: “The benefits, risks, burdens and effectiveness of a new intervention must be tested against those of the best proven intervention(s)”. In this case, the intervention involves giving mothers at risk of delivering a baby prematurely an injection of steroids. This steroid injection is widely regarded as one of the most effective interventions in perinatal care. It is standard of care to give mothers this injection in developed countries. Meta-analysis (the statistical tool for combining the results of many different studies) shows very clear benefit.[1] There have been strong campaigns to increase the use of this simple intervention, particularly in developing countries. Before the trial was published, Professor Joy Lawn, perinatal epidemiologist, noted that
“it is extremely unlikely, statistically, that antenatal corticosteroids would be shown not to work in African or Asian babies.”
There is no doubt that it would be regarded as unethical to perform a placebo controlled (or no-treatment controlled) trial of antenatal steroids in a country like the UK or the US.
So why the silence?
If the ACT trial had shown a reduction in infant mortality with antenatal steroids, I suspect that there would have been all of the negative attention that I described above – for authors, ethics committees, and journal editors. In fact, the ACT trial showed the opposite result to that seen in previous trials in developed countries. Babies who were in the half of the study that involved giving the steroid injection had a higher mortality than those receiving standard care. (The standard care in the countries where the trial took place did not include giving steroids) It was the steroid arm of the trial that had the 89 extra deaths. What is more, mothers who received steroids had a higher rate of infection.
It isn’t yet clear why steroids appeared to cause harm in African and Asian babies, when they have been clearly shown to benefit babies in developed countries. It may relate to higher rates of infection in those countries, or to differences in the health care setting and the availability of special medical care for sick newborn babies. It remains to be seen whether there are some mothers or some settings in low and middle income countries where steroids would be safe and effective.
However, what the ACT study makes abundantly clear is that it is dangerous to extrapolate medical evidence from trials in developed countries to very different low resource settings. Interventions that work in one setting may not work, or may cause net harm in the other setting.[2] It is for this reason that the ACT trial was ethical. Indeed, even if the study had shown clear benefit of steroids, the trial would have been ethical.
It is vital that patients in developing countries are able to access medicines and treatments that have been shown to benefit. It is crucial that they are not exploited by involvement in unethical research. However, it is also fundamentally important that they are not given medicines and treatments that have been inadequately evaluated, and that ethicists and ethical guidelines do not get in the way of carefully designed research trials that would evaluate them.
[1] The Cochrane collaboration is an international body for performing and disseminating systematic reviews of evidence. The logo for the Cochrane collaboration is based on the meta-analysis of trials of this very intervention.
[2] Another recent example in newborn care is the use of cooling for newborn infants deprived of oxygen at birth. While this has become standard of care in developed countries, trials in low income settings have not shown clear benefit, and indeed have shown apparent harm in some studies. See also http://fn.bmj.com/content/early/2010/11/09/adc.2010.184689.abstract