By Helen Petousis Harris 09/09/2020


Today AstraZeneca pushed the pause button on its late-stage trials of a COVID-19 vaccine. A clinical trial participant has experienced a serious health event and an investigation is underway to determine the cause. What does it mean?

A cautious approach – trials can halt to assess safety data

With over 200 COVID-19 vaccines now under development we are going to start seeing participants in the clinical trials have heath events. These events may or may not be completely unrelated to the vaccine under study. When you have what will ultimately be hundreds of thousands of people under observation in clinical trials, some will get sick, especially when you start recruiting older people into these studies. When this happens sometimes the trial will need to be halted while the event is investigated, and a lot of questions answered, briefly:

  • Did the person receive the vaccine or a placebo?
  • How long after the injection did this event occur?
  • What sort of health event is this? For example, is it heart attack? An infection? An injury? A seizure? Something else?
  • Is this event biologically plausibly explained by receipt of the vaccine?

There is more detail on these questions at the end of this blog.

What does the halting of a trial mean?

As the COVID-19 vaccine trials progress we are going to see periodic haltings of these trials. This is actually the second time the Oxford/AstraZeneca vaccine trials have paused. One reason for halting a trial is because a participant has had an adverse medical event. We must accept the possibility that one, or some, of these new vaccines may have safety problems and fail to progress further. We must also accept that if the event has been determined not caused by the vaccine the trial may resume. This is normal practice.

Who oversees the safety data on the trials?

Each clinical trial has a Data and Safety Monitoring Board (DSMB). This is an independent group of people who have expertise in the disease, the drug, statistics, or other relevant fields. While they will periodically review the study data for evidence of study-related adverse events, they can also make recommendations to the developer about continuing the study, modifying the study, or even terminating the study. When a serious event occurs they are alerted and will convene urgently. The data they review, and discussions are confidential.

In the case of COVID-19 vaccines, a large pool of safety-trained people have been recruited into a meta-DSMB. This team can liaise across all the individual DSMBs across all of the different vaccine products. This can help ensure the same level of rigour and standards are applied to all COVID-19 vaccines – at least those receiving support from CEPI (Coalition for Epidemic Preparedness Innovation).

The AstraZeneca COVID-19 vaccine

This vaccine was developed at Oxford University. Pharma giant AstraZeneca have partnered with Oxford to make this vaccine.  The vaccine is called a non-replicating viral vector vaccine and it uses an adenovirus that has had the gene for the SARS-CoV-2 spike protein inserted in to it. When injected, the vaccinee’s own cells will make the spike protein and the immune system generates a response. There is some experience with viral vector vaccines in humans and so far there are no particular safety concerns identified associated with this particular viral vector construct (ChAdOx1).

The adverse event

According to the NY Times the adverse event is a case of transverse myelitis (but this has not yet been confirmed by the company). FYI transverse myelitis is an inflammatory condition that affects the spinal cord. Biologically speaking many infections, such as the herpes and influenza viruses are known to cause this condition. It has even been reported associated with COVID-19. Historically there have been rare cases of such immune mediated inflammatory disorders associated with vaccines, but a firm causal link has never been established.

If this adverse event is confirmed as transverse myelitis I will update this page with a discussion.

What does this mean for COVID-19 vaccines?

This event does not affect other trials, particularly those using different vaccine platforms. It also demonstrates that safety is being taken very seriously. We need to be patient and let the trials run their course.

Some nitty gritty – assessing causality in more depth

Here is some more detail (based on the WHO causality assessment guidelines) about how these events are considered, in other words the sort of questions that need to be answered. First (obviously), the vaccine exposure must precede the occurrence of the event.

Depending on the type of vaccine involved, there may be definitive proof that the vaccine caused the event such as laboratory evidence, other times it may be classified as probable, possible, or coincidental.

  • Clinical or laboratory proof that the vaccine caused the event is most often found for live attenuated vaccines. (For instance, in a case of aseptic meningitis after immunisation with Urabe mumps vaccine virus, isolation of the Urabe virus from the cerebrospinal fluid is considered definitive proof that it caused the meningitis. Another example is isolation of the BCG agent from a focus of osteomyelitis.). In the case of subunit vaccines, such as HPV vaccines, we may accept that features consistent with a natural HPV infection may in part support a biological possibility.
  • Population-based evidence for causality – i.e. what is known about “Can it?” A definitive “yes” at the population level is consistent with causality at the individual level. A strong “no” at the population level is inconsistent with causality at the individual level. If there is no clear answer to the question at the population level, this will often lead to an indeterminate conclusion at the individual level. If there are significant numbers of individual cases, however, this clearly points to the need to try to answer the question at the population level. Here we would expect that when high quality studies compare the risk for the outcome between vaccinated and un-vaccinated they will show an increased risk among the vaccinated. For new vaccines still in trials this may not be possible.
  • Biological plausibility: In situations where the “Can it?” question has no clear “yes” or “no” answer, biological plausibility may provide support for or against vaccine causality. In other words, the association should be compatible with existing theory and knowledge related to how the vaccine works (see second bullet point).
  • Consideration of alternative explanations: In doing causality assessment on an individual case report, it must be remembered that in essence one is conducting a differential diagnosis. Thus, it is important to consider “coincidental AEFI” – i.e. an AEFI due to something other than the vaccine product, immunisation error or immunisation anxiety. All reasonable alternative etiological explanations should be considered, including:
    • pre-existing illness;
    • newly acquired illness;
    • spontaneous occurrence of an event without known risk factors;
    • emergence of a genetically programmed disease;
    • other exposures to drugs or toxins prior to the event;
    • surgical or other trauma that leads to a complication;
    • a manifestation of, or complication of, a coincidental infection that was present before or at the time of immunisation, or was incubating, but was not apparent at the time of immunisation.
  • Prior evidence that the vaccine in question could cause a similar event in the vaccinee. The concept of “re-challenge”, which is more commonly used in the assessment of causality in medicines, has been helpful for certain vaccine event considerations (e.g. Guillain-Barré syndrome (GBS) after tetanus toxoid vaccination, where GBS occurred on three separate occasions in the same individual within weeks of administration of tetanus toxoid).