The study by Gregory et al concluded that there was a statistically significant association between induction or augmentation of labor and autism risk in North Carolina for children born during the years 1990-1998 (diagnosed with autism between 1997 and 2007). In the letter, Miranda et al offered 3 possible explanations of the presumed association between autism and labor induction or augmentation; however, this is irrelevant because the data so far do not support that such an association exists.

The primary purpose of citing the paper by Gregory et al in our Journal editorial was to offer an alternative interpretation and a quite different conclusion to the same study. We pointed out that there was overdiagnosis of autism in the initial years of the study (prior to 2000) because of wrong wording in the Diagnostic and Statistical Manual of Mental Disorders , fourth edition regarding the diagnostic criteria for pervasive developmental disorder-not otherwise specified (PDD-NOS or atypical autism). This error was not given consideration or acknowledged by Gregory et al either in their article or in their subsequent news media interviews.

We also pointed out that according to the data by Gregory et al, when the correct diagnostic criteria for autism were applied after year 2000 (corresponding to birth years 1996-1998), there was no association between the augmentation or induction of labor and risk of autism.

In their letter, Miranda et al raised the unreasonable possibility that the misclassification in the diagnosis of PDD-NOS may have not have influenced their results if the children misclassified as PDD-NOS in the initial years of the study had been included in some other category of autism spectrum disorders. However, this is very unlikely because social impairment has always been the hallmark of autism, and it is a prerequisite for diagnosing any of the 3 individual autistic classifications namely autistic disorder, Asperger’s syndrome, or PDD-NOS.

As we stated in our editorial, instead of requiring “impairment in social interaction and in verbal or nonverbal communication skills,” the Diagnostic and Statistical Manual of Mental Disorders , fourth edition stated, because of an editorial error, that the “PDD-NOS category should be used when there is a severe and pervasive impairment of reciprocal social interaction or verbal and nonverbal communication skills, or when stereotyped behavior, interests, and activities are present.” Thus, children without social impairment who had problems only in communication or in stereotyped behavior, interests, and activities were classified as being autistic leading to an overdiagnosis of the condition; these children could not have been classified under any autism diagnosis in the study by Gregory et al because all 3 autistic conditions require the presence of social impairment for diagnosis. This leads to only one possibility, which is that those children misclassified as PDD-NOS did not have autism, thus leading to overdiagnosis of autism in the initial years of the study by Gregory et al.

Miranda et al also claim that the estimated association between labor induction/augmentation and autism adjusted out the potential influence of any yearly trend in autism diagnosis. However, we never questioned whether they did or did not adjust for year. Our claim was that their data regarding autism were based on an incorrect diagnosis for the birth years 1990-1995 and had nothing to do with statistical confounding.

We are pleased to see that when Miranda et al finally limited their analysis in the time period that the diagnostic criteria for autism were applied correctly (birth years 1996-1998), they found no statistically meaningful associations between augmentation or induction of labor and risk of autism. In our view, the authors of the article by Gregory et al should strongly consider retracting their initial conclusion regarding the association between induction or augmentation of labor and autism risk.

For the interested reader, a detailed point-by-point critique of the study by Gregory et al was addressed elsewhere.