The future of research into rotavirus vaccine

The future of a potentially lifesaving vaccine for developing countries has been imperiled by its recent withdrawal from the United States market. In August 1998, tetravalent rhesus rotavirus vaccine was licensed for routine vaccination in the United States on the basis of randomised controlled trials there and in Finland. The trials showed that the vaccine had an efficacy of 49-68% in preventing rotavirus diarrhoea overall and, importantly, 69-91% efficacy in preventing severe disease.^1–4

In July 1999 the US Centers for Disease Control and Prevention reported a clustering of cases of intussusception in the weeks after vaccination with tetravalent rhesus rotavirus vaccine, representing an additional risk of 1 in 10 000 for this complication.⁵ On the basis of this finding they recommended “postponing administration of tetravalent rhesus rotavirus vaccine to children,” and in October 1999 the manufacturer voluntarily withdrew the product from the United States market. This leaves researchers with a moral quandary: should randomised controlled trials of tetravalent rhesus rotavirus vaccine proceed in developing countries?

Firstly, it may be thought that other vaccines in development (for example, human-bovine rotavirus vaccine) may not cause intussusception. This is pure speculation. About one million children have been vaccinated with tetravalent rhesus rotavirus vaccine. Other vaccines in development have two or three orders of magnitude less experience in humans, numbers far too small to rule out a complication in the order of 1 in 10 000. Indeed, we cannot rule out the possibility that rotavirus itself causes intussusception in children who are predisposed to this problem; thus, vaccination with any live oral agent may be a triggering event.⁶ Secondly, the public health community generally believes that one must do no harm.

Assuming a worst case scenario of a 25% fatality rate from intussusception in developing countries, widespread use of tetravalent rhesus rotavirus vaccine could cause 2000-3000 deaths a year. For some, the prospect of causing this many deaths—or perhaps even any deaths—is morally untenable. The context of developing countries differs starkly from North America. Despite efforts to prevent death with programmes of oral rehydration therapy, about three million children die of diarrhoea annually.⁷ Of these deaths, approximately 600 000 to 800 000 are caused by rotavirus diarrhoea. Tetravalent rhesus rotavirus vaccine may prevent 80% of these deaths. If the next vaccine in development takes three to five years to get to the stage where tetravalent rhesus rotavirus vaccine is now, the choice to wait must be weighed against the cost of waiting: 1.4 to 3.2 million preventable deaths. Some have falsely assumed that inaction is a morally neutral state. But if one is culpable for vaccine related deaths, then one is also culpable for deaths caused by withholding the vaccine.

Is there a moral difference between a treatment that may cause a sick child to die and a vaccine that may cause a healthy child to die? Because public health doctors treat unhealthy populations rather than unhealthy patients the risk of death or serious disability must be lower with vaccines than with clinical treatments. The risks of tetravalent rhesus rotavirus vaccine seem comparable to the risks associated with measles, mumps, and rubella vaccine.⁸ The moral yardstick for the public health physician is ultimately the same as for clinicians: do the benefits of vaccination exceed the risks? In a developing country in which a child's risk of death from rotavirus diarrhea is 1 in 200 or greater the answer may well be yes.

Thirdly, as a result of the controversy over randomised controlled trials on the prevention of perinatal transmission of HIV in Africa and Thailand, there is sensitivity surrounding such trials in developing countries.^9–11 Some critics of the HIV trials believe that there should be a universal standard of care for a disease that is independent of the care that is available locally.⁹,¹⁰ Others argue more convincingly that this would lead to randomised controlled trials that are less responsive to the health needs of developing countries.¹¹

To make the standard of care in the United States the universal standard of care would only be unjust and would perpetuate the unjust distribution of healthcare resources globally. Tetravalent rhesus rotavirus vaccine was withdrawn from the market in the United States because the possibility of a 1 in 10 000 risk of intussusception seemed unduly high in comparison to only 20 deaths annually from rotavirus diarrhoea.¹² It is imperialistic to transfer this standard of care to a country in which 1 in 200 children die of rotavirus infection, and thereby to deny even further study of the tetravalent rhesus rotavirus vaccine.

Ethical randomised controlled trials must begin with an honest null hypothesis.¹³ Though the recent success of a randomised controlled trial of tetravalent rhesus rotavirus vaccine in Venezuela provides grounds for optimism, randomised controlled trials in other developing countries are needed because of various factors such as serotypical variance in rotavirus and the fact that the disease has an earlier onset and lacks the seasonality seen in developed countries.¹⁴,¹⁵ We don't know whether tetravalent rhesus rotavirus vaccine will be as effective in developing countries, but starting with an honest null hypothesis makes it ethical to proceed.