The number of cases of autism reported to state regional centers for the treatment of developmental disorders in California nearly quadrupled between 1987 and 1998, while the state's population was rising by only 20%. Diagnoses rose 16% between 1997 and 1998 alone. The state legislature commissioned epidemiologists from the University of California at Davis to study the problem.

Using clinic records, a diagnostic interview, and a questionnaire answered by parents, the researchers compiled information on nearly 700 children with developmental disorders in two age groups, one born in the mid-1980s and the other in the mid-1990s. About half of the children were diagnosed as autistic and half as retarded but not autistic.

The researchers did not find that parents and professionals had begun to use the label of autism more freely. The autistic symptoms of older and younger children were similar. About 20% of autistic children were misclassified as retarded but not autistic. That proportion was no higher in the older children than it was in the younger children, so misclassification could not explain the change. The upsurge in autism seemed to be real.

The CDC researchers surveyed children ages 3­10 in five counties of metropolitan Atlanta in 1996. They identified autism not only by referrals to clinics, as in the California study, but also by seeking other sources including children's hospitals, private physicians, and above all — 40% of the cases — school records.

They found nearly 1,000 children with autistic spectrum disorders, which included not only classic autism (two-thirds of the cases) but also milder related syndromes such as pervasive developmental disorder and Asperger's syndrome (normal intelligence and adequate language along with more or less severe autistic social and emotional deficiencies). That represented a prevalence of 34 cases in every 10,000 children — similar to the result of the California study and much higher than the rate indicated by earlier research. In studies conducted before 1985, for example, the rate of autism was about seven times lower, 5 per 10,000. Three American surveys in the early 1990s found similarly low rates.

But some skepticism is expressed in the CDC report and especially by an editorial commentator in the January 2003 issue of the Journal of the American Medical Association, where the results were published. The commentator argues that reports of autism could be rising because the public and professionals are becoming more aware of its existence and sensitive to its symptoms. More services are available, too. Federal law now requires states to provide special education programs for autistic children at public expense, and there is new evidence that some autistic symptoms respond to intensive behavioral treatment at an early age.

The California researchers acknowledge that the ability to notice and report signs of autism might have been improving. In California, among children born in the 1980s, 50% of autistic children were also diagnosed as retarded, compared with a much lower 22% among children born in the 1990s. Parents might have become more sensitive to autism in children who were not retarded.

Vaccination controversy

A popular suspect in autism is the measles-mumps-rubella (MMR) vaccine, which came into widespread use around 1990. Children are given the vaccine, which contains viruses that could theoretically infect the brain, at about 15 months of age. Among parents of the younger children in the California survey, more than 20% had avoided or delayed immunizing one of their children because of this fear.

To the relief of public health officials and parents concerned about contagious diseases, it is becoming increasingly clear that immunization is not the source of the problem. In the California study, developmental regression — reversal of previously normal intellectual and emotional growth, which should be seen if the vaccination were causing harm — did not occur more often in children born in the 1990s than in those born in the 1980s.

And northern European studies have been providing further evidence that vaccines are innocent. A Danish survey included all of the 500,000 children born in the country between 1991 and 1998. After adjustments for age, sex, birth weight, and the mother's education and social class, children who received the MMR vaccine — about 80% of the total — had a slightly lower rate of autistic disorders than those who were not vaccinated. And autistic symptoms were no more likely to appear shortly after vaccination than at any other time.

A survey of more than 500,000 Finnish children who received the MMR vaccine in the early 1980s confirms the Danish results. The number of children brought to hospitals for observation because of newly recognized autistic symptoms did not rise in the three months after vaccination.

The concern about vaccines results partly from the use of methyl mercury (also called thiomersal or thimerosal), a preservative that contains a small amount of the toxic heavy metal. It was removed from vaccines administered in the United States in 1999, but it is still used in other countries and regarded as safe by the World Health Organization.

A recent American study indicates that blood levels of mercury did not become unusually high in children who received thiomersal-containing vaccines for diphtheria, tetanus, whooping cough, hepatitis B, and influenza at 2, 4, and 6 months of age. They were compared with controls who took the same vaccines packaged without the preservative. Mercury levels were the same in both groups and well below the amount considered dangerous to health. The mercury in thiomersal also passed out of the body much faster than expected — a half-life of about one week rather than seven weeks — so it did not build up with repeated vaccinations.

These results should reassure parents of newborns facing vaccination. For now, no other satisfying environmental explanation for the rise in autism is evident. The University of California at Davis is establishing a research center to look for possible effects of pesticides, polychlorinated biphenyls (PCBs), heavy metals, and other contaminants. The CDC is tracking autism levels in several states using the same methods employed in the Atlanta research. Further study should tell us whether the rise in rates is in the eyes of the beholders or represents a true increase in the number of autism cases.