There has been so much discussion on the internet about autism and the role of MMR vaccines. Here at DKos alone, there are diaries
here,
here, and
here just for starters. Because of this, I thought I would post about an article that has just been published concerning regressive autism and the MMR.
Crossposted from OurSpectrum
Regressive Autism and MRR
There has been much debate on the
internets (snark intended) that regression in autism is due to the MRR.
A new study out headed by Catherine Lord, the grand-dame of autism research, seeks to test this hypothesis. It should be noted that regression itself is a confusing term. Does it truly mean a totally normal child that suddenly drops off a cliff? That is certainly one possibility. Another, and I am speaking right off the top of my head here, is that the parents failed to see developmental delay and that the child then stops
progressing at age 2-3. This does not jive with the stories of children who were outwardly loving and then suddenly become internalized. Nor, by research that shows children who are classified as regressing are more social than their counterparts who do not regress later.
So. What does the science tell us?
If regressive autism truly is a new, MMR-induced phenotype of the disorder, Fombonne and Chakrabarti argued, then one would expect, among other things, that children with regressive autism would have different symptom and severity profiles and a higher rate of GI abnormalities than children with ASD and no regression. If a particular phenotype of autism were associated with the MMR vaccine, then the age of onset for these children (with ASD and regression) would follow age at vaccination and would differ from the age of onset for children without regression.
The nice thing about this study is that they lay out very clearly what the hypothesis is (above) and then what the expected results would be if the hypothesis were true. The two questions to be addressed are:
1. Is there evidence for a `regressive phenotype' of ASD?
2. Is regression in ASD associated with the MMR vaccine?
The testable predictions are:
(1) The children with ASD and regression should have more social and communicative skills prior to loss than the children with ASD and no regression, but still show signs of atypical early development
(2) If children with ASD and regression manifest a new phenotype of the disorder, they should show different outcomes, in terms of social and communicative skills, than children with ASD and no regression
(3) If regressive autism is associated with GI symptoms, then children with ASD and regression should have a greater tendency to have GI disorders and/or symptoms than children with ASD and no regression.
(4) If regressive autism is associated with the MMR vaccine, then age at onset of autistic symptoms should more closely follow age at MMR vaccination for children with ASD and regression than for children with ASD and no regression.
The test of the hypothesis was based on the Collaborative Program for Excellence in Autism (CPEA). The CPEA draws on data from 13 total sites:
- Albert Einstein College of Medicine
- Boston University
- University of California - Irvine Medical Center
- University of California - Los Angeles
- University of Colorado
- University of Pittsburgh - Western Psychiatric Institute and Clinics
- University of Rochester Medical Center
- University of Utah - Utah Autism Project
- University of Washington
- Yale Child Study Center
- University of Chicago
- University of North Carolina
- University of Michigan
The patient pool was defined as those who were born in a Westernized county, lived in US or Canada from birth to age 3; received immunizations with records, at least 4 and no more than 15 when they received regression "interview", complete ADOS, AD-R, verbal and non-verbal IQ all within last 5 years. Children must have received diagnosis of AD, PDD-NOS, Asp and met criteria on ADOS and/or ADI-R. And/or? That really opens up the window doesn't it? Genetically identifiable diseases were excluded.
Word loss was used as the measure of regression. Children had to use at least three words other than mama and dada for at least a month, then stopped using all words before 36 months. Word loss appears to be the best predictor of regression and most memorable for parents. Also, those without word loss but with enough loss in other areas were included in the regression group (post-hoc?).
There is a ton of statistics and probability modeling in this paper. So, I'm going to skip over all of it, since it would require me actually learning all the math involved and go straight to the easy conclusions. These are the ones that really jump out at you.
Language - Verbal IQ measurements were interesting in that those without regression had a bell-shaped curve of distribution. Those with regression had a bimodal distribution, where a large percentage had lower Verbal IQ, but the same number had a high Verbal IQ (above 80) as in the non-regressive.
There are also a higher percentage of kids with regression that had high (bad) social reciprocity. Which seems a bit circular to me, kids with regression have low social activity by nature I assumed. The regression group had more children with both low VIQ scores and high ADI-R social reciprocity scores than the no-regression group. Again, to me this is just proving the obvious. Nothing wrong with that I guess.
The average age of onset for autistic symptoms was later for those children with regression. Remember that these children also have higher "typical" scores on tests of prespeech behavior, games, actions with objects, pretending to be parent, phrase comprehension, early vocabulary, and first communicative gestures. Quid pro quo? Higher functioning early, later onset, more substantial symptoms?
Given that regression has a later onset and that vaccination happened a the same time (statistically), you would expect that if MMR causes regression there would be a correlation between when you got the vaccine and when you regressed. I'm going to copy this whole paragraph:
Given that children with regression had a later age of onset of autistic symptoms than those without regression, but a similar age at MMR vaccination, one might predict that the two groups would differ in the timing of onset relative to vaccination; that is, a greater proportion of children with regression would have onset following vaccination than children without regression. In order to test this prediction, children with regression were compared to those without regression on timing of vaccination relative to onset. Table V shows the proportions of children in each group having onset before and after vaccination. Chi-square analyses yielded a significant difference between the groups in the proportions of children having onset before and after vaccination (χ 2 =10.07, df=1, p<.01). As predicted, a significantly greater proportion of children with regression had onset following vaccination, as compared to children without regression.
Table V. Age of Onset Relative to Age at MMR Vaccination
Regression (N=134) No regression (N=151)
Onset before vaccination 61 (45.5%) 97 (64.2%)
Onset after vaccination 73 (54.5%) 54 (35.8%)
Given this, do those with regression have onset of autistic symptoms closer to vaccination than those who do not regress? If MMR causes regression you would expect a significant number of regression kids to have onset near vaccination when compared to non-regressing.
BTW, note that in no regression 64% of kids have onset of autistic symptoms before vaccination.
Children in both groups who had onset after vaccination were classified according to the interval between onset and vaccination: 0-3 months, 3-6 months, and more than 6 months. Chi-square analyses found no differences between the groups in the interval between vaccination and onset. A t-test was done to determine whether there was less variability in the interval between MMR immunization and onset of ASD in the regression group, as one might expect if regression in ASD were associated with the MMR vaccine; again, no group differences were found.
There is one thing that raises some concern in my mind, because I am not a statistician. This graph:
The text says there is no difference in the proportion of children that are yet to show symptoms and the age past vaccination. However, that graph looks pretty divergent compared to:
This graph shows age at vaccination and you can see and believe that there is absolutely no difference between the two groups as far as when they were vaccinated (doesn't account for later onset of ASD in regression kids).
The conclusions then are:
If there is a `regressive phenotype' of ASD, then, it does not appear to be characterized by normal or near-normal early development.
...
children with ASD and regression had lower VIQ scores and higher ADI-R social reciprocity domain scores than children with ASD and no regression...a subset of the regression group had a relatively high verbal IQ, as indicated by a peak in the distribution close to 90.
...a greater proportion of the regression sample had high ADI-R social reciprocity scores relative to the no-regression group, {but} there was still a significant portion of the regression group that had moderate or low scores on this domain.
...parents of children with ASD reported higher rates of several GI symptoms than parents of children with ASD without regression.
...the present study provides no evidence that regression in ASD is associated with MMR vaccination
I know that this study will not lay to rest the MMR boogey-man. But, I think it should go a long way towards putting parents mind at ease that MMR is as safe as any vaccine and is not associated with regression or autistic onset within the population at large.