BMJ Volume 353, Number 9169 12 June 1999
Negative association between MMR and autism

Early last year we cautioned1 against premature acceptance of a hypothesis, proposed in a reported case-series, that measles, mump, and rubella (MMR) vaccination may cause autism, possibly by a mechanism involving induction of bowel abnormalities.2 We were especially concerned that the reported cases may have been due simply to temporal coincidence, and by the lack of supportive laboratory evidence. A special panel of the UK Medical Research Council also found the evidence unconvincing.3 Nonetheless, the hypothesis generated much media attention in the UK, with a subsequent drop in acceptance of MMR vaccinations.4

Less noticed were reports by the same group (and others) that highly specific laboratory assays in patients with inflammatory bowel disease (IBD), the postulated mechanism for autism after MMR vaccination, were negative for measles virus.5,6 This week sees the publication of two other reports that do not support a causal association between MMR (or other measles-containing vaccines) and autism or IBD. One report is by the Working Party on MMR Vaccine of the UK's Committee on Safety of Medicines.7 The Working Party was charged with the evaluation of several hundred reports, collected by a firm of solicitors, of autism, Crohn's disease, or similar disorders developing after MMR or MR vaccination. The Working Party conducted a systematic, standardised review of information from parents and physicians. Although acknowledging that it is impossible to prove or refute the suggested associations (because of variable quality of data, biased selection of cases, and lack of a control group), the Working Party concluded that the information available did not support the suggested causal associations or give cause for concern about the safety of MMR or MR vaccines.

In today's Lancet, Brent Taylor and colleagues provide population-based evidence that overcomes many of the limitations faced by the Working Party. Taylor and colleagues identified all 498 known patients with autism spectrum disorders (ASD) in North East Thames who had been born in 1979 or later, and linked them to an independent regional vaccination registry. ASD includes typical (core) autism, atypical autism, and Asperger's syndrome, but the results were similar when cases of core autism were analysed separately. The investigators first showed that the known number of cases of ASD cases has been increasing since 1979 and that there was no sharp increase after the introduction of MMR vaccine in 1988. Second, they found that, among affected individuals, the age at diagnosis was similar whether the child had been vaccinated before or after age 18 months, or had not been vaccinated, which indicates that vaccination does not result in earlier expression of autistic characteristics. Third, they showed that at age 2 years MMR vaccination coverage among the children with ASD was nearly identical with that in children in the same birth cohorts in the whole region, which provides evidence of an overall lack of association with vaccination.

Taylor and colleagues then used an innovative "case-series" method to assess the relative incidence of autism within predefined time periods after vaccination. These analyses involved three different measures of onset of autism (date of diagnosis, date of first parental concern, and date of regression) and two vaccine categories (MMR and any measles-containing vaccine). No statistically significant associations were found in the 14 comparisons, except for a slightly increased relative incidence (148) for the association of MMR vaccination and initial parental concern (which seems to have been due to parents' difficulty in recalling the precise age at onset and hence a preference for approximating the age as 18 months). Although the case-series method may be better suited for the study of an acute disease than of chronic disorders with an insidious onset, such as ASD, the results are buttressed by the lack of associations found in the other analyses.

The findings also are consistent with current understanding of the pathogenesis of autism, a syndrome defined by certain behavioural and developmental characteristics that may have a variety of causes. In few cases, however, is a specific cause identified. Autism has a strong genetic component, and associated neurological defects probably occur early in embryonic development.8 Thus, in most cases, autism represents a birth defect, although it may not be diagnosed until later in life when communication delays and characteristic behaviours become apparent. It seems unlikely therefore that a vaccination that is given after birth could cause autism.

Rare cases have, however, been described, of a normal child regressing and acquiring autistic characteristics. It is such cases of regressive disorders for which a biological link with vaccination is plausible.2 The onset of developmental regression tends to be clearly demarcated, making the disorder more amenable to the case-series method. Thus, Taylor and colleagues' analysis showing no association between vaccination and onset of regression provides especially persuasive evidence against the hypothesis that MMR may cause or exacerbate autism.

Taylor and colleagues conclude with the hope that their results ". . . will reassure parents and others who have been concerned about the possibility that MMR vaccine is likely to cause autism and that they will help restore confidence in MMR vaccine". Will the scientifically sound but essentially "negative" results published this week garner the same media and public attention as the initial report of the MMR-autism hypothesis? It is unlikely, as evidenced by the renewed media frenzy last week in response to another report by the group that proposed the hypothesis. This report was of an increased risk of IBD among individuals who had naturally acquired measles and mumps within 1 year of each other.9 The study had no data on MMR vaccine and the investigators specifically stated that they did not find a significant relation between monovalent measles vaccination alone and later IBD. Yet the popular media trumpeted the study as providing evidence that MMR vaccination may cause IBD. In such an environment it is critical to strengthen vaccine safety monitoring systems and risk-communication strategies to maintain public confidence in immunisations.

*Frank DeStefano, Robert T Chen


Vaccine Safety and Development Branch, National Immunization Program, Centers for Disease Control and Prevention, Atlanta, GA 30333, USA
1 Chen RT, DeStefano F. Vaccine adverse events: causal or coincidental? Lancet 1998; 351: 611-12.
2 Wakefield AJ, Murch SH, Anthony A, et al. Ileal-lymphoid-nodular hyperplasia, non-specific colitis, and pervasive developmental disorder in children. Lancet 1998; 351: 637-41.

3 Bignall J. UK experts convinced of safety of MMR. Lancet 1998; 351: 966.

4 Thomas D Rh, Salmon RL, King J. Rates of first measles-mumps-rubella immunisation in Wales (UK). Lancet 1998; 351: 1927.

5 Chadwick N, Bruce IJ, Schepelmann S, Pounder RE, Wakefield AJ. Measles virus DNA is not detected in inflammatory bowel disease using hybrid capture and reverse transcriptase followed by polymerase chain reaction. J Med Virol 1998; 55: 305-11.

6 Duclos P, Ward BJ. Measles vaccines: a review of adverse events. Drug Safety 1998; 19: 435-54.

7 Medicines Commission Agency/Committee on Safety of Medicines. The safety of MMR vaccine. Curr Probl Curr Pharmacovigilance 1999; 25: 9-10.

8 Rodier PM, Hyman SL. Early environmental factors in autism. MRDD Res Rev 1998; 4: 121-28.

9 Montgomery SM, Morris DL, Pounder RE, Wakefield AJ. Paramyxovirus infections in childhood and subsequent inflammatory bowel disease. Gastroenterology 1000; 161: 796-03.

Hosting by WebRing.