- Knowledgebase: Immunization Questions
- Questions about immunizations, side effects, and not vaccinating.
- 4. Measles-Mumps-Rubella Vaccine and Autism: The Rise (and FaIl?) of a Hypothesis - Top
- Measles-Mumps-Rubella Vaccine and Autism: The Rise (and FaIl?) of a Hypothesis
Jason L. Kastner, MD; and Bruce G. Gellin, MD, MPH
Pediatric Annals 30:7, July 2001
Two recent independent studies examining the hypothesis that the measles-mumps-rubella (MMR) vaccine is linked to the development of autism came to the same definitive conclusion-the hypothesis, derived from an initial observation that was presented in a small case series (n = 12), is not supported by the evidence.
Although two of the highest scientific courts (the American Academy of Pediatrics [AAP] and the Institute of Medicine) have come to the above conclusion based on thorough reviews of the evidence-published and unpublished, it remains to be seen whether the issue will be put to rest, or whether this unsubstantiated hypothesis will live on and become an urban legend in the folklore of medicine. In the 3 years since it first surfaced, the MMR-autism theory has received international attention and has been amplified by debates in the medical literature, professional circles, Internet chat rooms, and even state and national legislative bodies.
The most important setting in which this issue has surfaced is in the countless conversations between pediatricians and their patients' families. Reports from the pediatric front lines of parental and grandparental anguish and anger surrounding this issue suggest that questions and concern are likely to linger. No recent national surveys have specifically documented the proportion of parents who are concerned about the MMR-autism theory. However, a recent study found that more than 90% of pediatricians and 60% of family physicians reported that at least one parent had refused to allow his or her child to receive at least one recommended vaccine in the past year for some reason. An apparent increase in questions from parents and physicians directed toward the Centers for Disease Control and Prevention's (CDC's) National Immunization Information Hotline (CDC, personal communication, February 21, 2001) and an increase in requests to the manufacturer of the MMR vaccine for monovalent measles, mumps, and rubella vaccines (Merck Vaccine Division, personal communication, February 21, 2001) both suggest that the MMR-autism theory is part of the dialogue about immunizations.
MMR VACCINE AND AUTISM
At a time when vaccine-preventable diseases are at historic lows, the need to inform and edu cate patients and physicians about the scientific basis of our current immunization practices and policies has never been greater. Recent national surveys have documented that 1 in 4 parents of young children may have serious misconceptions about immunizations (eg, "too many vaccines will weaken my child's immune system"), and that up to one-third of family physicians and 12% of pediatricians do not recommend particular vaccines to parents routinely or occasionally.
The bottom line from both the AAP and the Institute of Medicine is clear-the combined MMR vaccine is not the cause of autism, nor can it be implicated in the apparent increase in the number of children being diagnosed as having autism spectrum disorders. By extension, there is no scientific or epidemiologic basis to support administering these three antigens separately
over time to prevent autism. Each of these reports provides a valuable synthesis of published and unpublished data that were made available for these comprehensive reviews and subjected to the principles of causality: strength of association, consistency, specificity, temporality, biologic gradient, plausibility, coherence, experimental evidence, and analogy.
Unfortunately, the science does not speak for itself. Although we can hope that the media, the Internet, and public deliberations will put their anecdotes into appropriate context and perspective, pediatricians can and must know the facts when they discuss these issues with patients' parents. The information age and the consumer movement provide ready access to a wealth of information from an array of sources. Most parentsÄoften armed with clippings, articles, and Internet printoutsÄlook to their pediatricians as the most important and trusted source of information and advice on childhood issues, particularly childhood immunizations.
Therefore, the purpose of this article is not to summarize these extensive reports, but rather to provide the pediatrician with an overview of the issue and a clearer understanding of the evolution of the MMR-autism theory. To be able to move forward, we must better understand how we got to this point.
THE GUT-BRAIN CONNECTION
The etiology of autism remains a medical mystery and is most likely multifactorial. Genetic, environmental, infectious, metabolic, and immunologic factors have all been investigated with
no clear-cut, single answer. According to the theory, put forth by Wakefield et al., autism results from a primary intestinal pathology whereby a leaky gut infected with measles allows the absorption of exogenous "gut-derived" opioid peptides from the intestinal lumen. It is then postulated that these unidentified peptides overwhelm the liver's filtering mechanisms and spill over to cross the blood-brain barrier and steer a genetically predisposed developing central nervous system on a path toward autism. Among the many unresolved issues in this theory is the fact that no abnormal gut-derived peptides have been detected in autistic children. Although such a "leak" may be plausible, why it would be in only one direction or so specific as to permit the passage of some molecules but not others of similar size, shape, and charge is unexplained."
The origins of the MMR-autism theory began with earlier investigations suggesting that "multifocal vasculitis" in the intestinal submucosa, perhaps due to persistent viral infection, is an early event in patients with Crohn's disease. Although there are several theories about the etiology of inflammatory bowel disease (IBD), a precise cause is unknown and the etiology is likely multifactorial (due to some combination of genetic, infectious, immunologic, and possibly psychological factors). The postulation that persistent viral infection could ultimately lead to IBD was fueled by a report from Wakefield et al. that measles virus had been identified in the intestinal tissue of individuals with Crohn's disease by electron microscopy, in situ hybridization, and immunohistochemistry.l5 However, other investigators failed to reproduce this finding. Rather, they isolated a cross-reacting intestinal antigen that was determined to be of human, not viral, origin.
Shortly thereafter, researchers published two epidemiologic studies that linked perinatal measles infection (defined as occurring during late pregnancy to 3 months of age) and in utero exposure to measles with the development of IBD. However, when they identified the study population, these researchers had prior knowledge that two cases of Crohn's disease had occurred in the offspring of women who were infected with measles during pregnancy. This selection bias limits the strength of the conclusions drawn by these studies. Another study published in the wake of these conclusions, suggesting a link between measles vaccination and IBD, drew significant national and international attention. In contrast, several other epidemiologic studies have consistently failed to show an association between either perinatal measles infection or vaccines containing measles virus and the subsequent development of IBD
O'Leary et al. reported isolating measles RNA in the intestinal tissue of autistic children with gastrointestinal complaints. The details of this study have yet to be published. However, when samples from O'Leary's laboratory were shared with an independent reference laboratory, the reference laboratory noted that ". . . there is questionable specificity as, on occasion, non-infected control tissues/cell have been recorded by O'Leary's laboratory as positive for measles virus" and "repeated tests of the same RNA were sometimes read differently by the O'Leary laboratory on sequential trials. Other investigators have presented data suggesting that paramyxovirus nucleic acid sequences were found in a portion of intestinal samples from patients with IBD, as well as a
smaller number of normal control subjects. One group identified measles RNA in peripheral blood cells from patients with IBD and autism. However, other researchers failed to demonstrate measles nucleic acid in IBD tissue using polymerase chain reaction. There is some evidence suggesting that measles virus may persist in unaffected tissue. However, presence does not necessarily imply pathology.
THE LEAP FROM IBD TO AUTISM: A VIRAL CONNECTION?
Although there is little epidemiologic evidence that children with autism experience a higher incidence of IBD, the connection between measles vaccination and autism rests on such a measles related abnormality in the intestine. The current MMR-autism debate came to the forefront with the publication and publicity of a small case series in 1998.2 In this report, Wakefield et al. described 12 children with intestinal symptoms and developmental regression. They observed that the onset of developmental problems was reported─by the parents or referring physicians─following the administration of MMR vaccine in 8 of 12 children. The average interval from vaccine “exposure” to symptom onset ws 6.3 days (range 1 to 14 days). However, the emphasis of their report was the endoscopic and histologic findings: of the 12 patients, 9 had evidence of ileal lymphoid nodular hyperplasia and 8 had what was described as "nonspecific colitis." On histologic examination, 7 children had ileal reactive lymphoid follicular hyperplasia without neutrophilic infiltrate or granulomata and all 12 had colonic lamina propria inflammatory cell infiltrates without granulomata. None of these endoscopic or histologic findings were seen in any of the 7 "normal" control subjects.
It was suggested, based on these findings, that MMR vaccination leads to an inflamed and dysfunctional intestine that results in autism, presumably via the hypothesized pathophysiologic process described earlier. Amplified by press briefings and extensive media coverage, this theory, portrayed as a medical alert, led to a measurable dip in MMR vaccine coverage in the United Kingdom.
INVESTIGATION OF A VACCINE SAFETY CONCERN
Concerned that media coverage of this proposed link among MMR vaccine, bowel disease, and autism could affect parental confidence in immunization and further reduce the use of the MMR vaccine, the United Kingdom's Medical Research Council reviewed the existing data and concluded that there was no evidence to support the proposed hypothesis.4l Subsequently, the United Kingdom's Medicines Control Agency established the Working Party on MMR Vaccine, and it also concluded that "the available information did not support the suggested causal association or give cause for concern about the safety of MMR or MR vaccines."
In a subsequent study, Taylor et al. performed an epidemiologic investigation into the proposed association between MMR vaccination and autism.43 Children with "autism" born in 8 health districts in the United Kingdom since 1979 were identified using special needs and school registers. The diagnosis of autism was confirmed in nearly 300 cases and these were then linked to immunization databases.
This study demonstrated a steady increase in the number of cases of children with autism (by year of birth) with no "step-up" when the MMR vaccine was introduced in the United Kingdom in October 1988. It also showed that vaccine uptake in children with autism was no different from that in children without autism and that there was no difference in age at diagnosis regardless of vaccination status or timing of vaccination. Finally, there was no evidence that regressive behavior clustered in the time intervals following vaccination.
Subsequent epidemiologic studies from Finland, the United Kingdom, and the United States demonstrated the same findings: a lack of correlation between the use of the MMR vaccine and new diagnoses of autism in the populations studied.264445 This evidence refutes the theory that the "epidemic" of autism in many countries followed the introduction of MMR vaccine. As the Institute of Medicine Committee concluded:
. . . the biological model linking MMR and [autistic spectrum disorders] ASDis incomplete and fragmentary. Possible immunologic and metabolic mechanisms have been described but have not been supported by validated and replicated controlled studies. While some believe that disrupted
viral immunity following administration of polyvalent vaccines could lead to atypical or persistent measles infection, possibly resulting in ASD or bowel disease, there is no biological precedent or sufficient evidence from existing research to support this scenarios
IS THERE AN EPIDEMIC OF AUTISM?
Autism was first described by Leo Kanner in 1943. However, the diagnostic criteria have changed over time and the criteria of the Diagnostic and Statistical Manual of Mental
Disorders, 4th edition, focus on three areas of impairment: social interaction, communication or language, and the display of restrictive or repetitive interests or behaviors. A diagnosis of "autis-
tic disorder" requires impairment in all three areas with onset in at least one area before the age of 3 years.46 Autistic disorder is actually a subtype in the larger category of pervasive developmental disorders, which includes Asperger's disorder, Rett's disorder, childhood disintegrative disorder, and pervasive developmental disorder not otherwise specified. The average age at symptom onset is approximately 18 months,47 shortly after the period when children receive most recommended childhood immunizations. For example, the MMR vaccine is recommended between 12 and15 months of age.
Although an apparent increase in the number of children being diagnosed as having autism spectrum disorders has been noted in many countries, less clear is whether this represents an epidemic. Wakefield and Montgomery argue that there is an epidemic of autism occurring in the United States, the United Kingdom, and elsewhere. A review of 8 epidemiologic surveys of autism between 1963 and 1983 estimated the prevalence of autism to be between 4 and 5 per 10,000. More recent studies estimated the prevalence of autism to be 7.5 per 10,000 and all pervasive developmental disorders to be 22 per10,000.51 Despite what appears to be a clear increase during that time, this may reflect a broader concept of autism, detection of cases with normal intelligence, changes in criteria, and increased detection due to better available services. Access to many special educational programs requires documentation of an appropriate diagnosis, so it is possible that this may contribute to the increasing number of children receiving this diagnosis.
The public debate about the epidemic of autism, and the hypotheses that it corresponded with the use of MMR vaccine, pointed to this report as supporting evidence. However, the data in this report have often been misrepresented as the incidence of autism in California. This report focused only on those served by the California Department of Developmental Services. Without an appropriate reference population (denominator), no conclusions can be drawn about changes in the incidence of autism over time. Factors such as population growth, migration flux, and changes in the birth rate make such interpretations difficult. Also, the diagnosis of autism occurs at younger ages, as reflected in the California data. Therefore, under these circumstances, arranging data by birth cohort can be misleading. Finally, the diagnostic criteria for autism have changed greatly during the past two decades. Thus, it seems likely that improving detection and expanding criteria to include more subtle cases that would have previously gone undiagnosed or been classified elsewhere account for at least a substantial proportion of the increases in children diagnosed as having autism in California.
SEPARATING ANTIGENS COMPARED WITH COMBINING THEM IN MMR VACCINE: DIVIDE AND CONQUER OR BETTER SAFE THAN SORRY?
Whereas the paramyxovirus theory of autism (mediated through intestinal pathology) hinges on measles virus (wild or attenuated), the suggestion that autism can be prevented by separating the MMR vaccine into its components (measles, mumps, and rubella) and administering each of these antigens separately over time stems from several lines of thought.
The first is that there is a transient period (weeks) of measurable but clinically insignificant immunosuppression following a measles infection. This finding is the basis of the recommendation to delay tuberculosis skin testing for at least 4 weeks following a measles infection or measles immunization to eliminate the possibility of a false-negative result. The recommendation to separate varicella and MMR immunizations, if not given concomitantly, by at least 4 weeks is also based on this transient immunosuppression. Although there are several theoretical mechanisms in which viral interaction could impair or affect the clearance of one or more of the vaccine strain viruses, the Institute of Medicine concluded that "there is no biological precedent or sufficient evidence from existing research to support this scenario."
The second comes from two European epidemiologic studies of early childhood exposures. These suggested that measles and mumps infections in the same year of life were risk factors for the subsequent development of IBD. The first study examined persons with IBD in Iceland born
between 1915 and 1984 who would have been younger than 6 years during epidemics of measles and mumps occurring in the same year. However, it is impossible to determine from these data
whether affected individuals experienced both diseases and experienced them during the same year. The second study examined persons with IBD using British cohort data from 1970. The data from this study relied on parental recall of how old the child was when he or she experienced several common childhood infections. Parents were asked this question when cohort members were 10 years old. Validation of the documented actual age of these infections was not performed.
The third is derived from ecologic studies. Although largely misrepresented, these suggest that the increase in autism diagnoses followed the introduction of MMR vaccine into national immunization programs. Before this, immunization had been based on monovalent measles vaccine.
The fourth is the theory that the earlier a person encounters a natural measles virus infection, the more likely he or she is to develop IBD. A recent examination of this theory showed "an indication"of a higher risk of IBD with younger age at measles infection; however, this finding was not statistically significant.59 The theory is also at odds with the global epidemiology of IBD because most early measles virus infection occurs in the developing world, where the incidence of IBD is lowest. Finally, although not specific to the measles, mumps, or rubella viruses, the "immune overload" hypothesis has been applied broadly to childhood immunizations. The anxiety stems from the belief that a developing infant's fragile immune system is somehow overwhelmed by the multiple antigenic load of an increasingly complex childhood immunization schedule. Although the evidence does not support the immunologic basis of such a theory, it is a widely held belief.
Although studies that have attempted to validate these various hypotheses have focused on measles virus (wild virus, vaccine virus, or both), the underlying assumption is that the combined MMR vaccine is somehow the "Trojan horse" that leads to autism. Thus, proponents of the vaccine connection theory of autism have proposed that delaying the administration of this vaccine until later in life, separating the combined vaccine into individual antigens, or both will somehow affect the host-virus interaction and thus prevent the initiation of the process that leads to autism. Given the fragmentary evidence supporting this model, the only impact of such a recommendation will be to make the current childhood immunization schedule even more complicated. The likely result of separating these three vaccines would be that an increasing number of children would not receive adequate immunologic protection against these three viral infections.
TOWARD THE ELIMINATION OF MEASLES: THE PARADOX OF EFFECTIVE PREVENTION
Vaccinations have had a tremendous impact on improving health globally. Measles immunization has decreased disease in the United States by more than 99% since the introduction of the vaccine in 1963. Currently, estimated vaccination coverage for one or more doses of MMR vaccine by 24 months of age is 88% nationwide. During the past several years, there have been fewer than 100 cases of measles in the United States each year. Most of these have been traced to importation from countries that continue to experience outbreaks of Measles.
Worldwide, measles still accounts for 36 million cases and 1 million deaths per year; most disease burden is in the developing world. Measles also accounts for 10% of mortality worldwide from all causes in children younger than 5 years. Indigenous measles was considered interrupted in the United States in 1993.64 However, an outbreak of measles occurred in the United States between 1989 and 1991. During this period, there were more than 55,000 cases and more than 125 deaths.65 The Netherlands experienced more than2,300 cases of measles and 3 measles-associated deaths during the past year. Ireland also recently experienced an outbreak of measles with more than 1,600 cases and 2 reported deaths. Coupled with international travel, the infectiousness of the measles virus makes it likely that there will continue to be periodic outbreaks, especially in populations with a large pool of individuals who remain susceptible.
However, because of the widespread use of vaccines containing the attenuated measles virus, measles is now an unusual childhood infection in most of the developed world. Yet, when it does occur, measles can still be severe. Respiratory complications include bronchopneumonia and croup. As a mucosal infection, measles can also acutely Prefect the gastrointestinal tract, leading to diarrhea, dehydration, malabsorption, and malnutrition. Acute central nervous system complications include encephalitis. The long-term consequence of subacute sclerosing panencephalitis has largely disappeered with the elimination of wild measles virus. Death, primarily due to respiratory and neurologic complications, occurs in 1 to 3 cases per 1,000.
CONCLUSION
Although recent independent reviews of published and unpublished research have rejected the hypothesis that the MMR vaccine causes autism, the public concern raised by this theory cannot be dismissed. Fear of the vaccine has led some parents to search for "measles parties"-actually seeking exposure to children with active measles infections in the hope that a natural infection would relieve them of the difficulty of deciding whether to immunize their child.
A better understanding of the willingness of parents to believe such a theory is beyond the scope of this article. However, the perception that this theory might somehow be true may cause some parents to operate on the principle of "better safe than sorry," regardless of the evidence. Because the science does not speak for itself, pediatricians should be familiar with this subject, prepared to discuss it with parents, and able to guide those who wish to do their own review of this hypothesis to credible resources. - Updated: July 11, 2001
-
[e-Mail me the Knowledgebase]- [Search
our Knowledgebase] - [Question Not Answered?]
|
