Editorial: An autism case series, vaccine hesitancy, and death by measles

Measles are back. Larger and more frequent measles outbreaks have been reported in 2024 in the United Kingdom and the United States, which predated the COVID-19 pandemic. Measles deaths worldwide rose to an estimated 136,000 in 2022, mostly children.

Immunizations have been one of the major contributor to the 20th century increased life expectancy alongside better nutrition, hygiene, and lifestyle, way ahead of medical technological prowess that keeps impressing us. Vaccination campaigns and other preventive policies (e.g. car safety belts and anti-tobacco campaigns) have saved and continue to save lives and reduce morbidity (remember tuberculosis or poliomyelitis) much more than da Vinci surgical robots or heart transplants. Vaccines are safe and post-licensure vaccine safety is continuously monitored with different overlapping population-based surveillance systems (Buttery & Clothier, 2022). Vaccines achieved the total eradication of smallpox in the late 1970s. A worldwide campaign to eradicate measles was well under way in the 1990s. Then, a case series published in a prestigious medical journal triggered fears of vaccine-induced autism in the public. Despite rigorous science rapidly dismissing the original claim, the measles-mumps-rubella (MMR) vaccine scare and fear of autism propagated. Years later, it has morphed into what we now refer to as ‘vaccine hesitancy’ (VH), MMR vaccine uptake is still below optimal levels, fueling ongoing measles outbreaks. How did we get there?

The saga started in 1998 with an editorial decision by the Lancet to publish a case series of 12 children suggesting a new syndrome of autism triggered by MMR vaccination had been discovered (Wakefield et al., 1998). The publication of a case series in the Lancet almost seems a contradiction in terms. As researchers and editors, we know how cautious we must remain before drawing causal inferences between two variables. There is an established hierarchy of research designs based on their relative strengths for causality assessment. Experimental designs are at the top (e.g. the RCT), followed by controlled observational (cohort and case–control) studies, and by much weaker ecological studies (relying on confounded correlations between group-level data); at the very bottom, lies the case series which, at most, helps generate hypotheses (especially when the knowledge base is quasi-inexistant) but is never sufficient to test causal ones.

While it is fair to assume that those who reviewed the 1998 manuscript and those who made the decision to publish it were unaware at the time of the fraudulent nature of the data (see: Godlee, Smith, and Marcovitch (2011); Deer 2011a, 2011b and Deer (2020)), multiple red flags were readily noticeable: its first author had spent previous years trying to prove that adult inflammatory bowel disorders were linked to measles vaccine, in studies that were unreplicated and dismissed by academic peers; having 12 children with autism ‘referred’ to an adult gastro-enterological surgeon was suspicious; no co-author had autism expertise; there was no control group of any sort; the MMR-autism connection entirely relied on unverified retrospective parental accounts asserting onset of autism days after MMR immunization in a previously normally developing child; regressive autism was presented as a novel syndrome (some Kanner's cases experienced developmental regression and loss of skills); implausible persistence of the measles virus in gut tissues. These accounts were unreliable and insufficient to speculate about a novel etiology of autism.

Despite numerous failures in the integrity of the peer review and editorial process (Deer, 2020), the predicted high impact of this article influenced the Lancet decision to publish. At JCPP, too, editors consider the potential impact of submitted articles as a criterion favoring acceptance for publication based on estimating the resulting increment in scientific knowledge. Instead, Wakefield et al. (1998)'s impact consisted of sensationalistic journalism, greatness of its author and institutional prestige, with disregard for the predictable negative public health consequences. By 2004–2005, the scientific community had conducted several independent population-based controlled studies that all failed to support a link between MMR vaccine and autism (I summarized them in Plotkin et al., 2020). Anecdotally, in one of these studies, we (with colleagues at the London School of Hygiene and Tropical Medicine) completed a large case–control study in England on the putative links between MMR vaccine and autism. It takes years to complete such studies and when in 2003 we submitted our manuscript to the Lancet, the journal (with the same editor) first returned it without review (‘it is no longer an interesting hypothesis’!). Our study had compared 1,294 autism cases to 4,469 matched controls and showed no association. This negative study was significant for public health but it was unworthy of media coverage. It took a persuasive phone call from one of us to challenge the initial rejection, have the manuscript sent out for review and finally published (Smeeth et al., 2004). The contrast between the editorial paths of these two manuscripts is striking and illustrates the vulnerability of the publication process that, in addition to the quality of the science, is influenced by undeclared or unknown personal, financial and social factors.

In 2004, 10 co-authors of Wakefield issued a personal retraction. The Lancet finally retracted the paper in 2010, 12 years later. In 2009, the Special Master of the US Vaccine Injury Compensation Court had issued a final opinion in the Omnibus Autism Proceedings about the lack of causal relation between MMR vaccine and autism in a class action. One hoped that the MMR-autism scare was finally over. However, as Jonathan Swift put it some time ago: ‘Falsehood flies and the truth comes limping after; so that when men come to be undeceived it is too late: the jest is over and the tale has had its effect’ (Swift, 1710).

MMR uptake has continued to be low in many countries, often lower than the 95% uptake level necessary to consolidate herd immunity. Even small % decrements in MMR uptake are associated with large increases in measles cases and rising financial health costs (Lo & Hotez, 2017). Fear of autism remains the first concern in the general public (Novilla et al., 2023), and in families within the autism community (Fombonne et al., 2020) resulting in MMR immunization delays/refusals in vaccinating younger siblings of children with autism (Zerbo et al., 2018). Vaccine refusal has existed since the invention of smallpox vaccine by Jenner in 1796. Dr. Moseley, an anti-vaxxer pioneer, warned of bestiality and human-cow hybrids (vaccine comes from the latin vacca that means cow), persuaded that ‘owing to vaccination the British ladies might wander in the fields to receive the embraces of the bull’. The profile of modern anti-vaxxers, often recruited among celebrities, has evolved and frequently contains features such as distrust of science and of expertise and an inclination for conspiracy theories (in a study of online information on the topic of autism and vaccine, Donald Trump was the name most often appearing on Web pages mentioning vaccine and autism (Arif et al., 2018)).

To describe the persisting reluctance to embrace life-saving, cost-effective, preventive measures (MMR or other vaccines), researchers now use the terminology of ‘vaccine hesitancy’ (VH) that appeared in the early 2010s and is an attenuated form of extreme anti-vaccine positions, equivalent to a broader anti-vax phenotype. WHO recently designated VH as one of the top 10 threats to be urgently combated. Yet, its contours are difficult to discern since it is not a unitary concept, rather it is a convenient descriptive term for a very heterogeneous group of people lying at the extreme of a distribution. Across VH individuals, variable constellations of arguments are found involving anti-vaccines sentiment, suspicion of evidence-based medicine, beliefs in the sanctity of mother nature, lack of science understanding, skepticism about experts, more extreme religious beliefs, distrust in government programs, low sense of social responsibility, a tendency to embrace conspiracy theories, fear of side effects (autism or else) together with actual decision/behaviors resulting in refusal or delay of immunizations. The latter behavioral part (vaccine refusal/delay) is what unifies VH individuals who otherwise became part of the VH subculture through multiple different pathways. Adding to this heterogeneity of motives, the focus of hesitancy varies according to vaccine type (HPV, MMR, and COVID-19), presumed adverse consequence (autism, auto-immune disorders), or vaccine ingredients (aluminum adjuvant, formaldehyde, thimerosal, etc.) (DeStefano, Bodenstab, & Offit, 2019). Here, we mostly confine our comments to the presumed MMR-autism links because of their direct relevance to mental health. A recent review of VH questionnaires emphasizes the measurement challenges that VH presents (Oduwole, Pienaar, Mahomed, & Wiysonge, 2022), owing to the diversity of reasons and foci among individuals with VH. Perhaps, separating more clearly the outcome of VH from its determinants will assist the optimization of these tools.

I became concerned that the use by researchers of VH terminology (including by me) has elevated VH to a higher status than it merits. By measuring and studying it, we substantiated it as a separate valid construct, a distinct phenotype of vulnerable subjects, almost amounting to a new personality or anxiety/stress syndrome that deserves special and delicate handling. As a result, health care providers give VH parents extra time and consideration, accept to post-pone an MMR immunization or to deliver its three components in separate shots (a practice known to contribute to incomplete immunizations), or now even train to motivational interviewing techniques to reduce ‘hesitancy.’ When vaccinations mandates have been relaxed (as in California or Texas), the generous distribution of non-medical vaccine exemptions for personal belief to VH parents has created community pockets of unvaccinated children from where outbreaks flared (Bednarczyk, King, Lahijani, & Omer, 2019). The direct and indirect costs of VH have been soaring for individuals and society.

A remarkable asymmetry exists. No equivalent attention is given to increased risks and costs incurred by non-VH parents as a result of VH, not only for their own vaccinated children but for the community where they live. Infants, immuno-suppressed children, children with vaccination failures, and those from underserved social groups with limited access to health care are unnecessarily exposed and at high risk of contracting measles. Yet, the costs attributable to VH are shared in full by those who followed vaccination recommendations. All public health programs have met with resistance by a minority, for example, restricting smoking in public places or wearing safety belts. However, one should note that there is no syndrome of ‘safety-belt ambivalence’ that could prevent the unattached driver to be fined; nor simply invoking the right to personal freedom will result in clemency. Likewise, smokers often have to pay a higher insurance premium. While coercion is not advisable, some form of personal accountability for VH might be considered.

In the United Kingdom, it took two decades for vaccination rates to recover. During that time there were 12,000 cases of measles, hundreds of hospitalizations, and several preventable deaths. Europe experienced a fourfold increase in measles cases and 35 deaths in 2017 alone. The incidence of measles in the United States was 0.03/100,000 in 2000 and rose to 0.39/100,000 in 2019. The year 2019 was the worst year for measles in the United States since 1992, with 1,282 cases reported, the vast majority among unvaccinated people. The forecasts for 2024 are worse.

Measles does not cause autism in children. But it kills them.