SEARCH

SEARCH BY CITATION

Autism, Asperger's syndrome, and pervasive developmental disorder–not otherwise specified (PDD-NOS) are conditions of early childhood characterized by multiple impairments in socialization and communication, as well as unusual interests and repetitive behavior. The central feature of PDDs is the significant impairment in social development. Differences in communication delay and repetitive behavior distinguish between the PDDs. For example, in addition to social impairment, autism requires the presence of language delay and repetitive behaviors or preoccupation with idiosyncratic topics (e.g., air conditioners, birds, birthdays of famous people, British royalty). Children with Asperger's syndrome do not have the history of impaired language development but are renowned for their restrictive interests. Children with PDD-NOS may have less profound language delay and may show less prominent repetitive behaviors and, thus, do not meet the criteria for autism or Asperger's syndrome. Psychotropic medications are commonly used in children with PDDs (Oswald & Sonenklar, 2007). To date, risperidone is approved for the treatment of tantrums, aggression, or self-injury in children with autism (Scahill et al., 2006), but there are no approved medications for the core features of the PDDs.

Recent events, including pronouncements about a link between autism and vaccines by presidential candidate John McCain during a campaign stop in Texas (Tapper, 2008), dramatic scenes in a prime-time TV show (Eli Stone, 2008), and a government settlement under the vaccine protection program (Marchione, 2008), have rekindled debate about the potential role of vaccines in the etiology of autism. The vaccine–autism connection turns on two proposals. The first involves the potentially toxic effects of thimerosal, which is an ethyl mercury–containing preservative used in selected vaccines in the 1990s (Blaxill, Redwood, & Bernard, 2004). The second involves the potential adverse effects of the combined mumps, measles, and rubella (MMR) vaccine (Singh & Jensen, 2003). Because these hypotheses rest on the conviction that the rate of new cases of autism has increased, this article, the first of three on the subject, considers the epidemiological evidence against an etiological role of thimerosal in PDDs.

Prevalence is a simple concept. It is the number of cases affected with a given condition divided by the population. It has no units and is usually expressed as a percentage or the number of cases per 1,000 or 10,000. Estimating prevalence is all about counting cases. There are three basic approaches in child mental health. The first method tallies all known cases from clinics, private practitioners, hospitals, and schools and uses that figure in the numerator of a prevalence statement. This method is likely to underestimate prevalence because children who have not yet come to a clinic and not been given a diagnosis will not be counted. Moreover, this method is vulnerable to spurious associations due to inequities in access to treatment. For example, based on cases ascertained in specialty clinics, early studies suggested that higher socioeconomic status was a risk factor in autism (see Fombonne, 2005). It is now clear that autism affects all socioeconomic levels, races, and ethnicities (Yeargin-Allsopp et al., 2003). The samples in earlier studies were biased because families of lower socioeconomic status did not have access to these specialty services.

To deal with this ascertainment problem, investigators may evaluate records in special education programs or hospitals for probable, although previously unidentified, cases. If resources permit, these probable cases undergo a clinical assessment in order to confirm the diagnosis of a PDD. Because this method includes previously unidentified cases, it will result in a higher, and more accurate, estimate of prevalence. The third, and arguably the best of the three methods, screens a large community sample in order to identify possible cases of PDD. Screened positive cases are then evaluated with diagnostic measures that classify children as cases or noncases. In PDDs, and in many other child psychiatric disorders, this method typically produces the highest and most reliable estimate of prevalence.

Prior to 1980, the prevalence of autism was estimated at 4 cases per 10,000 (Fombonne, 2005). Using better survey methods and broader diagnostic criteria, the prevalence of autism is now estimated at 20 per 10,000 children. The combined prevalence of autism, Asperger's syndrome, and PDD-NOS is also higher and is currently estimated at 60–65 children per 10,000 or approximately 1 in 160 children (Fombonne). (Note: The figure of 1 child per 150 for all three PDDs combined is frequently, although incorrectly, cited as the prevalence of autism.) Nonetheless, the apparent increase in the prevalence of the PDDs has fueled the ongoing controversy regarding the potential etiological role of vaccines in PDDs. If these increases in prevalence over time reflect a true rise in the rate of new cases, there would be a good reason to be alarmed and finding the explanation would be a public health priority. If, on the other hand, the apparent increase in prevalence is an artifact of systematic undercounts in previous estimates, then the claims about vaccine exposure provide more heat than light.

Findings from recent epidemiological studies conclusively demonstrate that exposure to thimerosal cannot account for the rise in prevalence of the PDDs. In 2003–2004, Fombonne, Zakarian, Bennett, Meng, and McLean-Heywood (2006) surveyed 27,749 children in 55 English-speaking schools in Quebec. To count cases, the investigators relied on identified PDD cases from the special education rolls in these 55 schools. Although cases were not independently evaluated by the research team to confirm the diagnosis, 155 of the 180 cases identified had been diagnosed at a regional medical center. The children ranged from 5 to 18 years of age. Thus, the children spanned across the era of low thimerosal exposure (those born between 1987 and 1991), to higher exposure (those born between 1992 and 1995), and no exposure (those born in 1996 or later when thimerosal was no longer used in Canada). If thimerosal exposure played a role in the etiology of PDD, the prevalence among those born after 1996 would be predicted to go down. The study identified a prevalence of 65 cases per 10,000 for all PDDs (21.6 per 10,000 for autism, 33.3 for PDD-NOS, and 10.1 per 10,000 for Asperger's syndrome). There was a steady upward trend in prevalence across time that was unrelated to thimerosal exposure. Indeed, there was a modest, but significant, increase in the likelihood of being diagnosed with a PDD in children born between 1996 and 1998 (after the Canadian ban on thimerosal) compared to children born between 1987 and 1995.

These results were replicated in a study conducted in the state of California (Schechter & Grether, 2008). This study used records from a statewide registry that tracks children receiving special education services. Records included a diagnosis of autism without differentiation between autism, Asperger's syndrome, or PDD-NOS for children born between 1989 and 2003. Once again, the study time period encompassed epochs of low, higher, and no exposure to thimerosal. Looking only at 3- to 5-year-olds for each year from 1995 to 2007 (children born between 1989 and 2003), investigators observed a steady rise in autism cases before and after the U.S. ban on the use of thimerosal in 2001. Specifically, from 1995 to 2003, the prevalence of autism in 3- to 5-year-olds rose from 6 to 30 cases per 10,000. From 2004 to 2007, the upward trend continued in 3- to 5-year-olds from 30 to 40 cases per 10,000. Merging these figures with the Quebec study is complicated by the fact that the California study reported the prevalence of PDDs without differential diagnoses. In addition, the California study reported prevalence for a specific age group (3- to 5-year-olds), whereas the Quebec study considered a wider age span of 5–18 years. Nonetheless, the trends in these two studies are remarkably consistent. The rise in identified cases of PDDs continued even after the prohibition of thimerosal.

The increase in prevalence of the PDDs has a threefold explanation. First, in 1994, with the release of the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM–IV), there was a broadening of the diagnostic criteria for autism. The DSM–IV also added criteria for Asperger's syndrome and clarified the criteria for PDD-NOS. Finally, better assessment methods have provided clarification across PDD diagnoses and improved the demarcation between PDD and non-PDD cases. Thus, more recent studies have overcome the systematic undercount of PDD cases in previous studies. The next paper in this series will examine the proposed toxicology of ethyl mercury. The third paper will examine the MMR hypothesis.

Acknowledgment

  1. Top of page
  2. Acknowledgment
  3. References

 From the Research Units on Pediatric Psychopharmacology Autism Network. This work was supported by a cooperative agreement grant from the National Institute of Mental Health: U10MH66764 (PI, L. Scahill) and by CTSA grant UL$$024139 (PI, R. Sherwin).

References

  1. Top of page
  2. Acknowledgment
  3. References