Does MMR Cause Autism?

In 1998, Wakefield and colleagues in London described 12 children with chronic enterocolitis and regressive developmental disorder. Ten of these children had autism, and in eight, the onset of regression was linked by the child's parent or physician to receipt of MMR. The authors suggested that replication of the three vaccine viruses in gut tissues caused a unique form of intestinal inflammation. They speculated that this led to the absorption of toxins that affected the brain, resulting in regressive autism.

Since a control group that was never exposed to the vaccine was not included in the study, a causal relationship between MMR and autism could not be established. Moreover, many questions about the integrity of the work were raised, spurred by the discovery that Wakefield had been commissioned by a group of lawyers to investigate a putative relationship between MMR and autism. At the very least, this raised the possibility that some of the patients in the 1998 study made their way to Wakefield because of his interest in this—in other words, the study population might have been biased toward patients who already thought their symptoms were brought on by the vaccine.

By 2004, these and other concerns prompted 10 of the 13 original authors to retract their previous interpretation of the study. In 2010, the United Kingdom's General Medical Council found evidence of serious professional misconduct by Wakefield and colleagues Walker-Smith and Murch; on February 2, 2010, The Lancet formally retracted the original paper and struck it from the published record. Unfortunately, this imperfect paper rekindled a new wave of antivaccinationism. The notion that MMR causes autism fed directly into the public's natural tendency to assume causality when two events are temporally associated—children get MMR at 1 year of age and the signs of autism become apparent right around the same time. MMR does not cause autism. Here is a summary of the evidence:

  • Persistent measles virus infection is not found in inflammatory bowel disease (IBD) or autism—Studies range from attempts to find the viral genome in bowel tissues using molecular amplification techniques to rigorous case-control studies. For example, one study of 30 patients with IBD, all of whom had had either natural measles infection or MMR vaccination, found no measles genomic sequences in bowel biopsies or blood lymphocytes using a sensitive nested PCR technique. A 2006 study found no differences in antimeasles antibody titers between 54 autistic children (51 had received MMR) and 34 controls (31 had received MMR), and no study subjects had detectable measles virus DNA sequences in their peripheral blood mononuclear cells. The final blow came in 2008, when investigators studied 25 children with autism and gastrointestinal disturbances and 13 children with gastrointestinal disturbances without autism. Ileal and cecal tissues were obtained and probed for measles virus sequences using molecular amplification techniques in three blinded laboratories, including the one wherein the original association between persistent measles virus and autism was reported. The results were unequivocal—only one case and one control had positive results, directly refuting Wakefield's hypothesis.
  • Case-control studies show no association between MMR vaccine and IBD—One such study done through the Vaccine Safety Datalink (VSD) showed no more risk of exposure to MMR among 142 cases with IBD compared with 432 controls without IBD.
  • There is no relationship between MMR uptake at the population level and the increased incidence of autism—Most studies show that autism cases are increasing, but many epidemiologists believe this is a reinfection of expanded case definitions, diagnostic substitution for other neurodevelopmental conditions, and more public awareness rather than a true increase in incidence. In any event, studies done in very different settings—the United Kingdom, California, and Montréal, to name a few—consistently show that the incidence of autism over time does not parallel the uptake of MMR, as it would if MMR caused autism. Another study from the United Kingdom showed no increase in cases of autism after introduction of MMR, which occurred in 1988. In addition, there was no clustering of cases of autism after receipt of MMR vaccine, even when the observation period was extended to many years.
  • There is no new form of gastrointestinal disease or autism that appears with the introduction of MMR—One study in the United Kingdom found that the proportion of children with developmental regression or bowel symptoms did not change significantly between 1979 and 1998, a period that included introduction of MMR. Similarly, a study of 262 patients with autism found no evidence for the emergence of a new form of the disease that included intestinal symptoms following widespread use of MMR. Other studies confirm that there is no distinct MMR-induced autism syndrome or “autistic enterocolitis.”
  • Case-control studies show no association between MMR and autism—One study done in the United Kingdom found that 78.1% of 1010 cases had received MMR before being diagnosed with autism, as compared with 82.1% of 3671 controls without autism, for an adjusted odds ratio of 0.86 (95% CI 0.68, 1.09). This means that the odds of being diagnosed with autism among persons who had received MMR were essentially the same as those who had not received MMR. Another study, done in Atlanta, showed there was no difference in the proportion of children vaccinated with MMR before 18 or before 24 months of age among 624 case children and 1824 matched controls.
  • Cohort studies provide strong evidence against an association between MMR and autism—Cohort studies are among the most rigorous epidemiologic investigations. The methodology is simple—a group of subjects is assembled, exposure or nonexposure to the risk factor is determined, and the subsequent development of the outcome is ascertained. The rate of the outcome is then compared between exposed and nonexposed persons; the ratio of the two is called the relative risk (RR), and an RR of 1 means there is no association between the exposure and the outcome. The beauty of retrospective cohort studies, wherein the cohort is identified in the past and followed to the present, is that the exposure and outcomes have already taken place, so that measurement of the exposure (in this case, receipt of MMR) cannot be biased by knowledge of the outcome (autism); therefore, robust inferences can be. The outcomes of autism or autistic-spectrum disorder were no more common in 1,647,504 person-years of exposure to MMR than they were in 482,360 person-years of nonexposure. The study further showed no association with age at vaccination, interval since vaccination, or the date of vaccination. This study provided very strong evidence against MMR as a cause of autism.

There is no scientific rationale for giving the monovalent components of MMR in lieu of the combination, and, in fact, as of 2008 the separate components were no longer available in the United States. Finally, the US Court of Federal Claims addressed the MMR-causes-autism theory as part of the Omnibus Autism Proceedings.

—Marshall, Gary S. "Addressing Concerns About Vaccines." The Vaccine Handbook: A Practical Guide for Clinicians. 3rd ed. New York: Professional Communications, Inc., 2010. 219-222. Print.

» Next article: Are Vaccines Made From Fetal Tissue?