Every clinical trial has exclusion criteria; rules about people having conditions or medical histories that may complicate results, but still, since it is not practical to completely genetically screen a large population, or to recognize any and all possible complicating factors, it often happens that people display symptoms that may not actually be involved with the drug or vaccine.
In fact, the approval of a drug will often weigh cost/benefits of known or discovered possible complications.
To stop a relatively large clinical trial for a single adverse event on an urgently needed vaccine suggests that there is a strong suspicion, but really no proof, that the adverse event, whatever it was, is a result of the treatment. It may, for example, be a cytokine storm or something of that nature, known from Covid-19. But it is possible to have this kind of immune response for other syndromes, after all people can and do die from viral infections that are much less lethal, overall, than Covid-19.
In an efficacy trial for a vaccine, one almost has to let the subjects out into the world where they will face a risk of infection. It would be unethical to deliberately expose people, so comparing a treated population with an untreated population is really the only way to test for efficacy. Therefore a patient could easily contract another disease, say a flu, to which they have a strong reaction.
I would be very surprised if this program was permanently halted based on my long experience with these kinds of trials.
Indeed, I have seen cases where I think pulling the drug for a series of known actual AE's may have lead to more deaths from the disease than the AE's would have.
Medicine is a statistical enterprise, and there are no definitive easy answers. I am pleased that Astra Zeneca is showing that they have not abandoned good science to rush into a positive result. It's a race, to be sure, and everyone is working hard. It reminds me of the time of AIDS before the approval of protease inhibitors.
= new reply since forum marked as read
