On March 30, 2021, The Atlantic posted my article on investigations in Europe into serious adverse events after people were vaccinated with AstraZeneca's Covid vaccine. It was titled, "We Need to Talk About the AstraZeneca Vaccine".

On April 5, Steven Salzberg posted a critique of it at Forbes titled, "No, We Don't Need to Talk About the AstraZeneca Vaccine".

I'd just urge anyone interested in this to read my article: it argues the case I want to make far better than this quick reply to Salzberg.

Early in it, I'd pointed to the blizzard of commentary that had been critical of European regulators for suspending the vaccine pending investigation. Much of it only referred to blood clots generally – which weren't occurring at elevated rates, as regulators emphasized. Commentators frequently did not, however, make it clear that the actions related to something else: rare clotting disorders that were co-occurring with low platelets – a combination that is far rarer still.

It turns out, Salzberg had, on March 22, posted a take along the lines I criticized. He expressed concern that suspension of the vaccine "plays right into anti-vaxxers' hands". He picks up from that post in critiquing my article. He writes mine was accurate, but "it also illustrated, in several ways, how not to communicate about vaccines". Salzberg points out that I wrote that investigations hadn't found an elevated risk of blood clots, and then goes straight to this:

"But then she writes that 'regulators must address the possibility (still unproved) that perhaps one in every 1 million vaccinated people could have a potentially fatal drug reaction.' This is just what not to write if you’re trying to communicate to the public about vaccines. Anti-vaxxers love to make claims about rare, unproven side effects, and then demand that scientists and governments must prove them wrong before anyone can trust the vaccines."

Discussion of the potential drug reaction I referred to is elided here completely. The end result is that Salzberg has now written 2 posts that fail to address the greater-than-expected rate of rare clotting disorders coupled with low platelets. I think my article makes it clear why I think skating over the problem leading to the suspensions isn't helpful, and that's for 2 reasons. It could lead to a loss of trust in the adequacy of vaccine safety monitoring, and treating this emerging condition in the usual way for blood clots could be catastrophic. As the British Society for Haematology says, "Clinicians need to be on alert for this syndrome, to understand how to make the diagnosis and to note the specifics of how to treat it".

Salzberg elides the second issue – the consequences for some people of not discussing an emerging clinical problem – and focuses only on the first. On that point, he writes:

"Here’s the flaw in that last statement: it assumes that anti-vaccine sentiment is currently very limited. That’s wrong. Reports in the U.S. indicate that 30% of Americans say they won’t get the vaccine. That’s a huge problem. Vaccine hesitancy, or mistrust, or whatever term you prefer, is already widespread. It’s unlikely to get 'far more severe' because it is already severe."

There are several issues to unpack here. For a start, it's looking at this only in a US context. As I stress, regulators in countries where this vaccine is in use are making decisions for their communities, and the context of the pandemic and vaccination options in those communities, as well as local values and expectations about response to investigation of safety concerns.

Let's take 2 of the countries I mentioned that have taken some form of action: Norway and Australia. Australia, the country I live in, has a population of over 23 million people, and roughly 70% of them – just over 16 million people – are adults under the age of 60. Only 20 people have died of Covid-19 in this community across the entire pandemic in that age range, with another 38 deaths in the 60-69 age group. We have only just started using the vaccine, which is being manufactured locally and is the mainstay of our vaccination program.

In Norway, one of the countries with excellent surveillance systems in Europe, the rate of the condition was estimated at 1 in 25,000 people under the age of 65, with a high fatality rate. If – and it remains if – anything like this were to occur at anything even close to that scale, then clearly it's something Australia would have to take very seriously, in the context of essentially zero risk from Covid at the moment. (We had 1 instance of the condition in the last few days in my state, in approximately 100,000 vaccinations.) For Norway, other vaccine options are greater than they are for Australia. And it's clear that the context in both these countries is very different from countries like Brazil, where thousands are dying of Covid-19 every day and relatively little of any vaccine is available right now.

Secondly, conflating hesitancy about Covid-19 vaccine with a ferociously "anti-vax" position is profoundly misleading. Even in the – outdated – single poll Salzberg links to for this statement, only 57% of those expressing concern about Covid-19 vaccination described themselves as opposing all vaccines. For the others, some didn't believe they needed the vaccine (they believe they have acquired immunity naturally, for example, or they are at too low a risk of exposure to warrant it). Others are concerned we don't know enough about the safety of these particular vaccines because of the speed at which they've become available. (There have been many polls and surveys of course – and the situation remains fluid.)

Salzberg argues it's unlikely to get "far more severe". I don't agree. People who are only hesitant – not opposed, just genuinely concerned that safety isn't being taken seriously enough – might not stay on the fence: they could go either way. Others, who haven't until now been concerned about regulators being diligent enough about safety, could become so.

Salzberg's final criticism of my article:

"Let me point out one final flaw in Bastian’s article. She explains that we can explain very low risks to the public, and as an example she states that 'an ordinary flu shot may cause a tiny number of recipients—just one or two per million—to develop an autoimmune disorder called Guillain-Barré syndrome'.

Actually, no. Studies have shown that there is no good evidence that the flu vaccine causes Guillain-Barré syndrome. This claim first arose in the late 1970s, when a swine-flu vaccine was reported to cause a very, very small increase in the risk of GBS. Reports of a supposed link between flu vaccines and GBS have never gone away, despite multiple studies that failed to link the two.

Bastian’s example proves how difficult it is to quash an erroneous story about the possible risks of vaccines: even after 40 years, an apparently non-existent risk of flu vaccines is still reported. My web searches turned up multiple ads for law firms that specialize in suing vaccine manufacturers over claims of harm caused by the flu vaccine."

While I get his frustration about claims about this syndrome, I think this approach is problematic. He truncates what I wrote, which is this:

"An ordinary flu shot, for example, may cause a tiny number of recipients—just one or two per million—to develop an autoimmune disorder called Guillain-Barré syndrome. The CDC describes this link as being 'variable and inconsistent,' but the slightly squishy facts are not concealed from the public."

The reason the CDC describes this as variable and inconsistent, is because there isn't just one influenza vaccine. The syndrome has been reported after vaccine, but certainly far from consistently. Salzberg's link for the statement that there is no good evidence of an association is for a single flu season (2012 to 2013). Even back in 2011 when the US National Academies studied the question of adverse events and vaccines, reviewed 21 epidemiological studies on this question alone (many of inadequate quality): there must be many more now. To pull out just one to claim it proves there is no association is problematic, even when it's done for a representative year. That National Academies report described the entire weight of evidence then as not proving it is an adverse reaction – but this next bit is key – "an association cannot be confidently ruled out, particularly for future vaccine strains". That's why I summarize this as "slightly squishy facts".

To cite just one study in support of a claim that there is no good evidence associating Guillain-Barré Syndrome with the flu vaccine, isn't ok – any more than it's ok to pick out one year where there was a mismatch between the flu strain that season, and the one that was predicted when vaccine was manufactured, and use it to claim there is no good evidence flu vaccines work. Cherry-picking is a communication practice that is risky in vaccine discussion in particular, when someone could sow doubt in people's minds by cherry-picking a couple of other studies to discredit you. Over-stating what we know about safety is a risk if we want people to trust what we say about vaccines.

Salzberg's final para:

Yes, we should take reports of possible vaccine side effects seriously. But no, we shouldn’t speculate publicly that the Covid-19 vaccines cause any particular harm before the evidence is in. The large numbers of vaccine-hesitant people in the U.S. demonstrate that once any rumor of harm gets out, even if it’s completely false, the damage can be irreversible.

That has little to do with the situation we are currently in. Several countries have chosen not to leave their investigations behind closed doors, so it's public, whether people like it or not. Last year, Salzberg wrote that he'd been wrong to previously encourage using these vaccines without waiting for phase 3 trials. "Phase 3 can identify less-common side effects that might still be very bad", he wrote. "These less-common side effects are a big risk of moving too quickly." The thing is, phase 3 trials can still only identify relatively common side effects, and the risks to public trust are still there. In an interconnected world, once one country goes public, whether or not to discuss it publicly is a moot point. The question becomes, how do we do it? Because as Salzberg points out, people who want to dissuade people from using the vaccines will be on it. When there's so much effort going into misleading people from every direction, and so much effort going into judging regulators for their actions, I believe we're better off well-informed about why this is happening.

I agree with Salzberg, that if after all this is, we're sure this is not a vaccine reaction, we have a tough communication road ahead, while the road for malign actors is easy. But it became easy for them the moment some countries starting taking action. I don't think leaving people poorly informed is the best way to start off down this road, regardless of the conclusion. And telling people drug regulators are making bad decisions is no way to help.

Hilda Bastian

April 6, 2021


Update, soon after posting: After objection by Salzberg, I changed a sentence. It previously read "To pick out one study, and make a claim that it proves Guillain-Barré Syndrome cannot be caused by the flu vaccine, isn't ok...and use it to claim that proves flu vaccines don't work". It was updated to: "To cite just one study in support of a claim that there is no good evidence associating Guillain-Barré Syndrome with the flu vaccine, isn't claim there is no good evidence flu vaccines work."