Book Review: Vaccines, Part 1
Vaccines are like drugs? Hit me.
Vaccines are like drugs, according to eminent medical methodologist, anti-establishment gadfly, and exiled Cochrane whistleblower Peter C. Gøtzsche in Vaccines: Truth, Lies, and Controversy (Skyhorse Publishing, 2021). And, after reading this book, I want more. As soon as possible. Straight in my arm. This effect was unexpected on my side, unintended on his, and may be catching.
This post summarizes what I learned and loved, and why Gøtzsche accidentally convinced me to get myself more shots as soon as possible. The next post turns to critique.
After a quick summary, I sketch Gøtzsche’s key methodological contributions — attending to potential vaccine interactions, non-specific effects, a denominator error that pervades vaccine discourse, and identifying certainty bias. Then, I turn to how this work changed my own thinking about particular adult and pediatric vaccination choices, before closing with Gøtzsche’s strongest bigger-picture points on opportunity costs, perverse incentives, and how vaccine debates involve broader social dynamics (“us versus them”) and value choices (liberty versus security).
Summary
Gøtzsche’s core message is nuanced, counters hype on both sides, and is written in a highly accessible fashion. He writes “everyone should get vaccinated against measles” but “no one needs to get vaccinated against influenza” (p. 145). On this spectrum, he puts HPV vaccination in the middle, expressing concerns about possible serious neurological harms/deaths — along with Covid vaccination, which kids and some countries don’t need, protection from may not last, and from which long-term possible harms are unknown (p. 187).
Overall, he invites critical thinking rather than compliance. It’s a reflection of how polarized the current discourse is that it seems exceedingly rare to hear an expert speak openly about vaccine costs and benefits alike, and about vaccines with mostly beneficial risk profiles (measles) and deleterious ones (flu). But he does, and it helps clarify the evidence.
Above all, I really enjoy Gøtzsche’s signature speaking truth to power, and his emphasis on human dignity and reason — especially in the face of the us-versus-them mentality that fear tends to provoke in contexts including this.
Gøtzsche’s main point, in haiku summary:
Vaccines are like drugs —
we must weigh benefits and
risks, and learn much more.
Key methodological contributions
Vaccine interactions
Gøtzsche writes “We know far too little about the harms caused by simultaneous use of several vaccines and about the harms related to their sequence when not given at the same time” (p. 88). He cites at least one case where giving one vaccine along with another attenuates the response.
This is also one of several points where his humility and humor make reading fun, as when he proclaims: “This is terribly interesting. I needed to write a whole book on vaccines before I found evidence that vaccines can interact negatively” (p. 202).
Non-specific vaccine effects
What’s the right outcome measure for vaccine efficacy? Is it whether you get the disease? Whether you pass it on? How severe your case is? Whether you’re hospitalized? Whether you die from it? What about whether you die — from anything?
Wait, what?
All-cause mortality is a great outcome variable, because we can count it (ostensibly) even in places where it’s harder to get data due to extreme poverty and corruption. But why would we expect it to make sense in the context of vaccines, treatments intended to prevent specific diseases?
Because, apparently, vaccines may also have non-specific effects — effects unrelated to preventing the target disease, like improving (or worsening) general immune function.
Not so surprising, considering that they’re immune system-modulating therapies, often with long-term (potentially lifetime) actions. (Though two are recommended, one childhood dose of live measles vaccine may confer lifelong protection.)
(If you’re wondering how we can tell and why looking for this isn’t just a recipe for healthy volunteer bias or something, the answer is randomized trials.)
Some of these non-specific effects look bad: the combination DTP (diphtheria, tetanus, pertussis) vaccine used in poorer countries may heighten female all-cause mortality, while the TDaP (tetanus, diphtheria, acellular pertussis) formulation used in wealthier countries is less well-studied for non-specific effects; both are non-live. Other non-live vaccines (e.g., for polio) may similarly heighten female all-cause mortality.
We don’t know why non-live vaccines may increase female all-cause mortality. But, in retrospect, it looks like we should have been keeping tabs on that outcome variable for all vaccines since the start of vaccination. Oops.
(And it’s at least prima facie plausible that these sorts of effects are gendered. We know women’s immune systems work differently, making them substantially more vulnerable to most autoimmune disorders.)
The good news: some of these non-specific effects look good. As leading vaccine researchers Christine Stabell Benn et al recently wrote on Sensible Medicine:
The NSEs are most pronounced while a given vaccine is the most recently administered, making the sequence and combination of vaccines very important… We have found incredibly beneficial effects of some (live) vaccines, and less positive, even harmful, effects for females of other (non-live) vaccines.
So, potentially all you have to do is have a live vaccine to top off the non-live vaccines, and that may reset your immune system so the net effect is an immune benefit instead of whatever decrement is causing the all-cause mortality increase from non-live vaccines.
There are, however, downsides to approaching the problem this way…
Every vaccine carries risks, and those risks are counter-balanced by fewer benefits when you take one just for possible non-specific effects.
Pregnant women are supposed to get vaccinated against whooping cough (aka pertussis) in the second or third trimester to protect the baby. This is usually only available in a combination form with other non-live vaccines, such as TDaP. They can’t then get a live vaccine top-off, since it might harm the fetus. But it’s specifically females who appear to be at risk from non-specific vaccine harms. This raises the question of whether current public health guidance pushing for universal pertussis in pregnancy may cause net harm. It would certainly incur less risk to vaccinate just for pertussis; but that shot seems to be generally unavailable, because its market would be smaller and its profit margin less.
We don’t know why this should work.
We don’t know that this does work.
Logistics: Unless you’re a doctor or best friends with one, you’ll have to socially engineer how to get the treatment you think you need (as always). Explain the medical literature and hope for the best? Shop around with different doctors, with different incentives (GP, gyn; public system, private clinic)? Hit a travel clinic with a simple story about planning travel? These options reflect a pay-to-play medical system where patients are tasked with both understanding the medical literature well enough to know what’s in their interests, and finding a doctor whose incentives align well enough to give it to them.
Methodology: Isn’t this whole thing based on outcome-switching? Isn’t that bad?
We still haven’t run large-scale randomized trials to estimate possible non-specific effects of common vaccines, to try to figure out what combinations/sequences are best (if it matters). Gøtzsche calls for this, noting the public would have to pay for them so that there aren’t perverse corporate incentives, but that they needn’t cost much and would likely pay for themselves in avoiding harms and showing us how to accrue benefits.
He might have then restated that this literature, this hole in it, and this need aren’t new, and so the situation doesn’t look likely to change to let us get this data anytime soon. This ties into research on research (e.g., Altman on the need for “less research, better research, and research done for the right reasons”). He doesn’t get into that level; maybe it would be a different book?
Anyway, in his discussion on non-specific effects, Gøtzsche cites two relatively recent PubMed-indexed papers by some of the same leading researchers who cowrote the Sensible Medicine essay (p. 47). Here are a few money quotes from both of them:
In “Developing the concept of beneficial non-specific effect of live vaccines with epidemiological studies,” Peter Aaby and Christine Benn recount the modern history of non-specific vaccine effects, which were suspected already from the start:
When smallpox vaccine was introduced around 1800, bacillus Calmette-Guérin (BCG) against tuberculosis in the 1920s and oral polio vaccine (OPV) in the 1960s, there were suggestions that these live attenuated vaccines reduced mortality more than expected. However, scientific follow-up was limited and the concept of beneficial NSEs did not become mainstream (Clin Microbiol Infect, 2019 Dec;25(12):1459-1467).
In their research, they observed all-cause mortality decreases following live measles virus vaccination in poor countries in the 1970s. “Subsequent observational studies and randomized trials confirmed beneficial NSEs of smallpox vaccine, BCG [for TB] and OPV [for polio].” In their view, “Observational and experimental research has shown beneficial NSEs of four live attenuated vaccines: smallpox vaccine, BCG, OPV and MV [measles].” They call for vaccine schedule planners to take these effects into account.
This conclusion contrasts with Gøtzsche’s call for large-scale randomized trials. It seems that the experts on NSEs already think there is enough evidence to base public policy on it, at least as opposed to the current status quo of ignoring it while mass vaccinations continue. Why?
They sort-of address this question in the other relevant article Gøtzsche cites, aptly entitled “How to evaluate potential non-specific effects of vaccines: the quest for randomized trials or time for triangulation?” (Christine Stabell Benn et al, Expert Rev Vaccines, 2018 May;17(5):411-420; open source). Benn et al argue RCTs are costly and risk introducing their own biases, and so they should be triangulated with observational studies with different biases to form a better causal model. (Which implies that we should also go ahead and triangulate with the randomized trial and observational data that we have… Since there is no perfect data waiting for us at the end of the perfect trial.)
So the disagreement doesn’t seem to be over whether we should run RCTs or not (we should), but rather over what we should do in the meantime (Benn et al: use what we know now to possibly improve vaccine schedules; Gøtzsche: everybody read the literature yourself, and try to avoid vaccine co-administration and sequence a live vaccine last). Benn et al want the burden at the system level, while Gøtzsche wants it to remain at the patient level until the evidence base is bolstered with large-scale randomized trials specifically designed to answer related questions.
The Sensible Medicine essay itself cites this slightly more recent Benn et al article, “Vaccinology: time to change the paradigm?” (Lancet Infect Dis, 2020 Oct;20(10):e274-e283). It suggests vaccine reform could substantially lower child mortality by harnessing positive non-specific effects, avoiding negative ones, and attending to possible interactions (e.g., with vitamin A supplementation).
The German child vaccination schedule already implements the suggestion to ensure infants obtain their measles vaccination after their TDaP, so that a live vaccine comes last in the sequence. Probably parents everywhere do have to watch out for this for themselves, even though that’s obviously not a burden that should be put on them.
Denominator errors
When I was in grad school, I wrote a long essay involving calculations of the risk of flu complications versus the risk of Guillain-Barré syndrome from vaccination for someone with lupus, because my mom had lupus and I was trying to advise her (without giving medical advice, of course). The apparent problem was that the risk of both flu complications and vaccine complications, such as the autoimmune overreaction thought to cause Guillain-Barré, is higher for people with autoimmunity. But it turned out that the heightened risk was comparatively greater from the infection than from the vaccination. (Don’t cite me on this; the essay would have been written around 2009, and I’m not digging it back up to check references and see how this literature has since evolved.)
My usual editor for this sort of piece at the time, who usually loved my stuff, declined to publish it. He was afraid that, even though I got the “right” answer in the end (recommending vaccination as less risky than infection), just asking this sort of question could stoke vaccine hesitancy, validating non-compliance with directed public health behavior.
It’s a good thing he didn’t have a spine, because it turns out I made a denominator error. The one Gøtzsche pinpoints in the broader vaccine discourse (p. 88): Comparing Guillian-Barré after flu versus after vaccination takes incomparable denominators, because more people get vaccinated than get the flu if the vaccination recommendation is universal, but the infection is not. This applies generally, since the U.S. Centers for Disease Control and Prevention (CDC) recommends everyone get the flu vaccine, but not everyone gets the flu.
In other words, we can’t compare a certain risk from opting to get a vaccine that carries possible risks, with an uncertain risk from possibly getting (but possibly avoiding) an infection. Or, if we do, a correct such comparison has to recognize that any risk incurred from the vaccination has a different denominator (100% of those vaccinated) than the risk incurred from going unvaccinated and possibly getting the infection (not 100% of those going unvaccinated).
Calling out certainty bias
Gøtzsche is good about calling out certainty bias. For instance, he rebukes the European Medicines Agency (EMA) for saying the HPV vaccine benefits outweigh harms, when we don’t know (p. 185).
Personal takeaways
The only reason reading this book convinced me to get more instead of fewer vaccines is that I was already fairly well-informed — and had thus stopped getting flu shots and Covid vaccines. It convinced me to fight my co-parent harder to stop flu vaccinating our son, because that appears to incur net risks. And it convinced me to get pneumococcal vaccines for my recurrent sinusitis, and an MMR booster hoping to get some myself of those sweet, sweet non-specific effects. Allow me to explain…
Pediatric flu vaccination looks net risky
I was aware that evidence for flu vaccination is flimsy at best, and had asked my ex to let our four-year-old forego it as a result. But he’s pro-vaccine, doesn’t read the relevant scientific literature himself, and generally favors obeying doctor’s orders. The pediatrician recommended the kid get the flu shot this past fall, when I was pregnant and didn’t have the energy to fight it. I tried. I failed. The kid got the shot.
Having read this book, I won’t let that happen again.
Gøtzsche writes:
As noted in Chapter 1, studies by Canadian researchers indicated that people who received a seasonal influenza vaccine in 2008 had an increased risk of getting infected with another strain in 2009. They replicated their findings in five different studies. Other researchers have reported that annual influenza vaccination hampers development of CD8 T-cell immunity in children” (p. 89, footnotes 93 and 94).
“Someone saved my life tonight”
One of the places where his interpretation was different than mine, but it was hugely valuable that Gøtzsche put out information that enabled me to make my own call: He mentions pneumococcal vaccines that are CDC-recommended for adults over the age of 65, but that he himself (being over 65) won’t get, because he doesn’t find the evidence of benefit compelling, while the evidence of possible harms is non-trivial (p. 200-1). But he also mentions the jabs in connection with sinusitis, since it’s a common infection caused by the targeted Streptococcus pneumoniae (p. 199).
Somewhere around the middle of my fourth round of antibiotics and steroids in the past year for recurrent sphenoid sinusitis due to autoimmunity, reading this turned on a little light in the birdhouse of my soul. It suggested that these vaccines might help me get off this particular carousel. There is little that could improve my quality of life more. I will be forever grateful to Gøtzsche for this accidental suggestion, provided following through on it doesn’t kill or maim me.
Hypothetical future possible problems
Being totally done with whatever the repeated pregnancy TDaPs may or may not have done to my already subpar immune function, I’m on the hunt for a doctor who will give me an MMR booster for the possible non-specific effects. This raises the obvious hypothetical future possible problem: What do I then do if I manage to actually find my life partner and have at least one or two more babies after getting boosted?
Then I get the TDaP again in late pregnancy to protect the baby from pertussis. And then I’d need to pick one or more additional live vaccines to top it off.
(Not that I have thought this through, but my current live vaccine for non-specific effects hit list goes: (1) measles — best evidence for NSEs, (2) chickenpox — maybe NSEs explain that dementia effect, (3) tuberculosis — best remaining risk-benefit balance among live vaccines. My current life problem list above those working solutions for this very hypothetical dilemma: (1) take care of existing babies, (2) build meaningful career, (3) find life partner.)
Big-picture insights
Attending to opportunity costs
Often omitted from otherwise credible cost-benefits assessments, resource reallocation or opportunity costs should play a role in our net effects estimates for vaccination among other interventions.
Gøtzsche reminds us to consider these costs. For instance, he notes it would be way more effective to spend resources “preventing young people from starting smoking rather than convincing their parents that their daughters and sons should get vaccinated against HPV” (p. 180).
Gøtzsche’s trademark: speaking truth to power
Gøtzsche is widely known as a critical-thinking gadfly who speaks truth to power. In many places, his trademark accuracy and clarity shine through, as when he calls psychiatrists’ response to lockdown challenges “devastating” (p. 134) and denounces screening for mental disorders along with psychiatric medications.
In another example, Gøtzsche denounces the WHO for putting the same people in charge of approving a vaccination program as it does in reevaluating that recommendation — a clear conflict of interest (p. 194-5). And also for putting DTP coverage 90% as a goal globally by 2015 (3 doses), when it’s not proven to increase child survival, and that should be the program performance indicator (p. 195). This is especially worrisome in light of Aaby’s findings suggesting, rather, “that the DTP vaccine likely increases total mortality in low-income countries” (p. 195). So he suggests we need trials where kids receive a live vaccine or vaccines (he suggests MMR) earlier after DTP (non-live ones), to see how early is safe and if it improves outcomes.
Us versus them
The herd protects, and the herd endangers. Gøtzsche groks this duality, though he could grapple more openly with its tensions.
Protection: The origin of the word vaccination is the Latin word for cow (vacca), because Edward Jenner — British physician, scientist, and early vaccination practitioner involving cowpox for smallpox — named cowpox “Variolae vaccinae”
(p. 206-7). We want the good kind of herd immunity (against measles through vaccination), but not the bad kind of herd behavior (e.g., bandwagoning against dissent; mindless going-along when evidence is — as usual — exegetical and incomplete).
In his HPV vaccine chapter, he recounts unfair attacks (including by the EMA) on Danish researcher Dr. Louise Brinth from the Danish Syncope Unit in her research on dysautonomia, chronic fatigue syndrome, and encephalomyelitis in patients who presented with post-vaccination symptoms (p. 146-7). On these attacks, he remarks “the mentality of the Inquisition will be with us forever” (p. 148). He sees them as part of a larger problem of the vaccine context triggering “us versus them” psychology, although he doesn’t appear to explicitly recognize why this makes sense: fear primes right-wing authoritarianism, typically making us more strongly identify with our in-group and fear/loathe the out-group(s).
Similarly, in describing media misreporting on his PhD student Lars Jørgensen’s HPV vaccine harms dissertation, he writes, “It was like religious intolerance, following the well-known confession for vaccines, ‘are you for or against,’ that the three examiners were described as vaccine heretics, which they are not. And the research was not at all controversial; it was of high quality” (p. 177-8).
And, describing his reaction to the flawed and incomplete Cochrane HPV review: “I felt it would be appropriate to change Cochrane’s motto, trusted evidence, into touted nonevidence, or, as Tom Jefferson phrased it, garbage in, garbage out, with a nice little Cochrane logo on it” (p. 173). (Having had my follow-up comment on their breastfeeding support review recently censored, I heartily concur.)
He continues: “There is no doubt that my criticism of the Cochrane HPV vaccine review played a major role in my expulsion from Cochrane” (p. 175).
And notes the opposition responded to logical critique with accusations of postmodern confusion: The National Board of Health’s director, Søren Brostrøm, “talked about alternative facts when researchers were skeptical of the work done by the drug companies and EMA. He launched the idea that people are opposed to the HPV vaccine because we live in ‘a postfactual society.’ Denigrating remarks like these can have the opposite effect of the intended one” (p. 179).
Conclusion
Before I read this book alongside its companion, Brian Deer’s The Doctor Who Fooled the World: Andrew Wakefield’s war on vaccines (see review), I accidentally went to the playground with my ex and our son without bringing my reading. I was excited about these books, and suggested they should be translated into a prime-time version and taken door to door in response to the crisis of confidence in MMR following Wakefield’s fraud… And not just that mistake, but the series of larger establishment mistakes including requiring risky, unnecessary Covid vaccination for healthy young people in the U.S., not fixing vaccine safety reporting problems and instead continuing to push HPV vaccination in spite of unresolved concerns, and the general fallibility and non-neutrality of humans doing science and running scientific institutions.
His response was the same as Brostrøm’s: To say that, then, everyone would have their own facts, and we don’t want that. We just want compliance.
It’s a remarkably internally inconsistent position coming, as it usually does, from a left-liberal political position that tends to accuse the Trump administration and other conservative political groups of enacting authoritarian policies with disregard for human rights. The difference is whose authority is being used to violate whose liberty: “ours” or “theirs.”
This book is Gøtzsche’s attempt to indirectly counter that stance by translating enough scientific evidence for a popular audience that we can make up our own minds as reasonable people. In that way, it’s a love letter to the liberal democratic principles of individual autonomy and limited, legitimate state powers which underpin modern, Western civilization. An effort to “show not tell” the argument that consent can work, because reason can prevail. We can have adult conversations about complex topics with life-and-death, real-world consequences. We can have nice things.
Opponents of this classical liberal vaccine stance tend to misunderstand the nature of the underlying evidence, which is typically (like most scientific evidence) complex, ambiguous, and changing. We have to keep learning and thinking critically about it, in order to keep making informed choices. No rest for the wicked; that means us.
This faith in people as reasonable, and civilization as sustainable on (or, indeed, only through) a bedrock of trust in that reason, makes this book worth reading. Even if it may be wrong, at least in parts…