A Review of Hviid et al.’s 2019 MMR-Autism Study

  

Q: Why didn’t Physicians for Informed Consent (PIC) use Hviid et al.’s MMR-autism study published in 2019 in the Measles Vaccine Risk Statement (VRS)?

A: Hviid et al.’s 2019 study is weaker than the study analyzed in the Measles VRS for the following reasons:

1) Hviid et al.’s study has a smaller proportion of unvaccinated children.

The study discussed in the PIC Measles VRS included 96,648 unvaccinated subjects out of 537,303 total subjects, about 1 in 5.5 subjects. Hviid et al.’s study includes 657,461 subjects, but only 31,619 were unvaccinated, about 1 in 20.8 subjects. Hviid et al.’s study is 4 times weaker in its representation of the unvaccinated population.

2) The statistical power of Hviid et al.’s study is still not enough to prove that MMR causes less permanent harm than measles.

Hviid et al. found an “adjusted autism hazard ratio of 0.93 (95% CI, 0.85 to 1.02),” which means that they did not rule out the possibility that the vaccine increases the risk of autism by a factor of 0.02. The unvaccinated risk of autism found in the study was 525 out of 31,619. Multiplying this by a factor of 0.02 is about 1 in 3,000. That is 3 times greater than the 1 in 10,000 chance of dying from measles. As is usually the case in these kinds of studies, “no increased risk for autism after MMR vaccination” was found within the confines of the statistical power of the study. But that power is not enough to detect the incidence of autism occurring in an amount as small as the risk of fatal measles.

3) The final results of Hviid et al.’s study are biased based on the age of diagnosis.

The study states, “The mean age at first autism diagnosis was 7.22 years (SD, 2.86), and the mean age among autistic disorder cases was 6.17 years (SD, 2.65),” yet the results cited in No. 2 above include children only followed until 3 years of age. The study states, “Ending follow-up at 3 years of age yielded a slightly lower aHR (0.73 [CI, 0.53 to 1.00])” and “Ending follow-up at…10 years of age produced…0.97 [CI, 0.87 to 1.07].”

Therefore, the study did not rule out the possibility that the vaccine increases the risk of autism by a factor of 0.07 in children followed up for 10 years. Among unvaccinated children who were 10 years or older at the end of the study (i.e., those born between 1999 and 2004), the study found 418 autism cases out of 15,876. Multiplying this by a factor of 0.07 is about 1 in 540. That is 18.5 times greater than the 1 in 10,000 chance of dying from measles.

4) The small unvaccinated group in Hviid et al.’s study was at an unusually higher autism risk.

The study computed a “disease risk factor” to gauge which children were possibly predisposed to autism regardless of vaccination status. Among the vaccinated group, 61,296 of 625,842 (10%) subjects had high risk scores. By comparison, 4,465 of 31,619 (14%) unvaccinated subjects, 1.4 times more subjects than the vaccinated group, had high risk scores.

Figure 3 shows that if only children with very low autism risk scores are included in the study, the RR is 0.93 (CI, 0.74 to 1.16), and the study did not rule out the possibility that the vaccine increases the risk of autism by a factor of 0.16 in children at very low risk of autism. Among unvaccinated children at very low risk of autism, the study found 91 autism cases out of 7,590. Multiplying this by a factor of 0.16 is about 1 in 520. That is 19 times greater than the 1 in 10,000 chance of dying from measles.

5) The great majority of the unvaccinated group in Hviid et al.’s study started to vaccinate but then stopped.

Of the 31,619 subjects that were not vaccinated with MMR, 26,890 (85%) had been vaccinated with a different vaccine before and declined MMR after that. The study did not rule out the possibility that a significant portion of those children may not have been vaccinated with MMR because of a severe reaction to a previous vaccine or because of a health condition that discouraged them from being vaccinated. Figure 3 shows that if only children who had not been vaccinated with DTaP-IPV/Hib are included in the study, the RR is 1.09 (CI, 0.77 to 1.56), and the study did not rule out the possibility that the vaccine increases the risk of autism by a factor of 0.56 in children not vaccinated with DTaP-IPV/Hib. Among unvaccinated children not vaccinated with DTaP-IPV/Hib, the study found 64 autism cases out of 4,729. Multiplying this by a factor of 0.56 is about 1 in 130. That is 77 times greater than the 1 in 10,000 chance of dying from measles.

Conclusion: Although Hviid et al.’s study provides compelling evidence that MMR is not causing autism in 1 in 100 cases, it has provided no evidence that MMR is not causing autism in more than 1 in 10,000 cases (more than the case-fatality rate of measles). To the contrary, the study did not rule out the possibility of MMR causing autism in 1 in 3,000 children overall, in 1 in 540 children followed for a 10-year period, in 1 in 520 children at low risk of autism, and in 1 in 130 children who had a similar vaccination history. 

Terms

aHR: Autism hazard ratio

CI: Confidence interval

DTaP-IPV/Hib: Diphtheria, tetanus, acellular pertussis – inactivated poliovirus/Haemophilus influenzae type B

MMR: Measles, mumps and rubella

RR: Relative risk

SD: Standard deviation

To assess the risks of measles compared to the risks of the MMR vaccine, visit https://physiciansforinformedconsent.org/measles.