Arguably, the genesis of the most recent iteration of the anti-vaccine movement dates back to 1998, when a remarkably incompetent researcher named Andrew Wakefield published a trial “study” in the Lancet that purported to find a link between “autistic enterocolitis” and measles vaccination with the measles-mumps-rubella (MMR) trivalent vaccine. In the wake of that publication was born a scare over the MMR that persists to this day, 11 years later. Although peer reviewers forced the actual contents of the paper to be more circumspect, in the press Wakefield promoted the idea that the MMR vaccine either predisposes, causes, or triggers autistic regressions. Even though over the next several years, investigations by investigative journalist revealed that not only was Wakefield’s research funded by trial lawyers looking to sue vaccine manufacturers for “vaccine injury” when he did his research (for which he is now being ), but during the Autism Omnibus trial in PCR technology showed that he was incompetent. Even worse for Wakefield, in February 2009 Brian Deer published a based on strong evidence that Wakefield may very well have falsified data for his Lancet paper.
None of this mattered. Andrew Wakefield still enjoys a cult of personality among the anti-vaccine crowd that no revelation seems able to dislodge, even the revelation that at the time he was both in the pay of trial lawyers and working on his study, Andrew Wakefield was also for a rival measles vaccine. Indeed, the anti-vaccine propaganda blog Age of Autism bestowed upon him last year its “” as the “persecuted” scientist supposedly fighting for truth, justice, and anti-vaccinationism against the pharma-funded or brainwashed minions of the “.” In the meantime, MMR uptake rates in the U.K. have plummeted over the last decade, far below the level needed for herd immunity, to the point where, last year the Health Protection Agency in the U.K., 14 years after the local transmission of measles had been halted.
Since Wakefield’s study was released, a number of studies have shown that there is no epidemiologically detectable link between vaccination with MMR and autism, including one by a researcher who once appeared to be a believer in the idea that vaccines are somehow linked with autism, Mady Hornig. Hornig actually tried very hard to replicate Wakefield’s 1998 Lancet study, only this time with more children, and she found no link between MMR and autism using methodology similar to Wakefield’s. None of these studies has had any effect on the anti-vaccine movement, except to motivate them to circle the wagons even more, as J.B. Handley of Generation Rescue did when he launched a website called , whose purposes are to launch fallacious and pseudoscientific attacks on studies failing to find a link between vaccines and autism (often involving accusations of being a “pharma shill”), to promote the lousy science that gives the appearance of supporting the hypothesis that there is a link between the MMR vaccine and autism, and then slime anyone who points out how deceptive their attacks were.
Now, yet another study has been released studying whether there is a link between MMR vaccination and autism. Yet another study has failed to find a link between MMR vaccination and autism. Yet another study is all set to be attacked by Generation Rescue and the anti-vaccine movement. The sad and sordid history of reactions of the anti-vaccine movement to studies that do not support its belief in the unsinkable rubber duck of a myth that vaccines cause autism. This study was published online in The Pediatric Infectious Disease Journal by a group from Department of Epidemiology and Preventive Medicine, Jagiellonian University, Collegium Medicum, Krakow, Poland (a Polish group, my people!) and entitled . It’s yet another nail in the coffin of the myth that the MMR causes or contributes to autism. Indeed, this study not only shows that MMR vaccination is not associated with autism but that it may even be protective against autism. True, for reasons I will discuss shortly, I doubt that that latter interpretation is true, but there’s no doubt that this study is powerful evidence against the view that there is an association between MMR and autism. Unfortunately, I fear that all the nails in my local Home Depot would not be enough to keep the zombie of this pseudoscience from rising from its grave yet again.
Here’s the abstract of the paper:
OBJECTIVE: The first objective of the study was to determine whether there is a relationship between the measles-mumps-rubella (MMR) vaccination and autism in children. The second objective was to examine whether the risk of autism differs between use of MMR and the single measles vaccine.
DESIGN: Case-control study.
STUDY POPULATION: The 96 cases with childhood or atypical autism, aged 2 to 15, were included into the study group. Controls consisted of 192 children individually matched to cases by year of birth, sex, and general practitioners.
METHODS: Data on autism diagnosis and vaccination history were from physicians. Data on the other probable autism risk factors were collected from mothers. Logistic conditional regression was used to assess the risk of autism resulting from vaccination. Assessment was made for children vaccinated (1) Before diagnosis of autism, and (2) Before first symptoms of autism onset. Odds ratios were adjusted to mother’s age, medication during pregnancy, gestation time, perinatal injury and Apgar score.
RESULTS: For children vaccinated before diagnosis, autism risk was lower in children vaccinated with MMR than in the nonvaccinated (OR: 0.17, 95% CI: 0.06-0.52) as well as to vaccinated with single measles vaccine (OR: 0.44, 95% CI: 0.22-0.91). The risk for vaccinated versus nonvaccinated (independent of vaccine type) was 0.28 (95% CI: 0.10-0.76). The risk connected with being vaccinated before onset of first symptoms was significantly lower only for MMR versus single vaccine (OR: 0.47, 95% CI: 0.22-0.99).
CONCLUSIONS: The study provides evidence against the association of autism with either MMR or a single measles vaccine.
Indeed it does. It’s worth talking a bit about the methodology of the study. First of all, this is a case control study, which means that it’s retrospective and therefore not randomized. On the other hand, it was a case control study by Sir Richard Doll that was the first outside of Nazi Germany (whose scientists, oddly enough, had found evidence linking smoking to lung cancer more than two decades before the Surgeon General issued his report in 1964) to find an association between tobacco smoking and lung cancer, a finding that was subsequently followed up in cohort studies and found to be valid. Be that as it may, case control studies, instead of prospectively following a population over time, study a population of patients who already have a disease or condition and then try to identify factors associated with the development of that condition. Basically, this involves picking a control population that is equivalent to the study population, and this is how the investigators did it:
Subjects were identified using general practitioner records in the Lesser Poland (Małopolska) Voivodeship in Poland. The sample population of this study included children aged 2 to 15 years diagnosed with childhood or atypical autism, classified according to ICD 10-criteria as F84.0 or F84.1, respectively. Every diagnosis of autism was made by child psychiatrist. Dates of these diagnoses were recorded in general practitioners files. Cases with uncertain diagnosis of autism, secondary to disease state or trauma, were excluded. Two controls were selected for each affected child, individually matched by year of birth, gender, and physician’s practice. The first 2 children visited the physician after the time of the autistic child visit who met entry criteria served as controls.
This is a reasonable strategy to use to pick the control group, as it means that the children would be coming from the same pediatric practices as the autistic children and would be about as comparable as it is possible to make them in this sort of trial design. Another strength of this study is that it used physician reporting for vaccination history and the age at which each autistic child was first diagnosed, rather than relying on parents’ reporting, which is prone to serious confirmation bias. True, the parents were also asked when they first suspected their child’s symptoms might be related to autism, and their knowledge and beliefs regarding the cause of autism, but vaccination was not mentioned in order to avoid biasing the parents’ answers. Cases of autism were then analyzed and considered to have been vaccinated if they received the MMR prior to the onset of autistic symptoms. Controls were considered vaccinated if they had received the MMR before the age of onset of their matched case controls. Finally, the authors corrected for other potential risk factors for autism, including mother’s age, education, gestation time, medications during pregnancy perinatal injury, and the APGAR score of the child at birth. Those that appeared significant in univariate analyses were then subjected to multivariate analysis.
The money tables are Table 3 and Table 4:
What’s surprising about these results is that the multivariate analysis found a relative risk (RR) of autism for children vaccinated with any vaccine before the onset of autistic symptoms of 0.65 and 0.28 for children vaccinated before their “official” diagnosis of autism. When looking at the single vaccine for measles or the trivalent MMR vaccine, both appeared to be protective against autism, but the MMR appeared to be considerably more protective than the single shot measles vaccine. (This comparison could be made because the Polish national health service only covered the single dose measles vaccine and not the MMR until 2004, meaning that before 2004 parents who wanted the MMR vaccine had to pay extra.) Indeed, this study evne found that for children vaccinated with MMR before their diagnosis of autism the RR = 0.17, suggesting a six-fold decreased risk of autism in children vaccinated with MMR!
Is this really true? Is MMR vaccination really that protective against autism? Probably not. An effect of that magnitude would very likely been picked up in one or more of the large population-based studies that failed to find a correlation between vaccines and autism. There is also one shortcoming in this study is that it only looked at autistism and not other pervasive developmental disorders; however there is ample other evidence that MMR is not associated with PDDs other than autism, and the original claim of the anti-vaccine movement was that MMR causes autism. So how to explain such a result? One possibility is simply random chance, given that the sample size, although reasonable, is not that large. Another possibility is that there is a confounder that wasn’t adequately controlled for. Whatever the case, here’s one thing to remember about retrospective studies in general. They often find associations that later turn out not to hold up under study using prospective studies or randomized trials or, alternatively, turn out to be much weaker than the retrospective study showed. They do not so often find a result that is exactly the opposite of hypothesis tested for. In other words, when such a study is used to look for a positive association between a factor and a specific condition or disease, it is quite uncommon to find a negative association, particularly one this strong. One possible reason for this seeming protective effect observed was discussed by the authors:
The decreased risk of autism among vaccinated children may be due to some other confounding factors in their health status. For example, healthcare workers or parents may have noticed signs of developmental delay or disease before the actual autism diagnosis and for this reason have avoided vaccination.
I have speculated before that a study of “vaccinated” versus “unvaccinated” children could be similarly confounded by parents who have one autistic child, become convinced that vaccines caused it, and therefore don’t vaccinated subsequent children that they might have. The main point to remember is that, even though this study is not compelling evidence that MMR is protective against autism, at the very least, the Polish study is strong evidence against a positive correlation between vaccination with MMR or single vaccination against measles and the development of autism. It is yet another pebble in the mountain of evidence that vaccination with MMR is not associated with autism or “autistic enterocolitis.”
It’ll be fun to see how long it is before Handley adds this study to his “” website and renames it “Fifteen Studies.” Of course, at the rate studies failing to support Generation Rescue’s cultish clinging to the unsinkable rubber duck of a belief that vaccines cause autism, Handley will have to rename his site every few months at least, and either I or one of my partners in crime here at SBM will have to slap down yet another scientifically ignorant attempt at discrediting each new study.
On and on it goes. Meanwhile, autistic children pay the price for quackery related to anti-vaccine beliefs, and normal children face the resurgence of vaccine-preventable diseases, thanks to the efforts of anti-vaccine cranks like J.B. Handley, Jenny McCarthy, Jim Carrey, Barbara Loe Fisher, and the rest of the vaccine denialists.