On 7 February 2018, the National Vaccine Injury Compensation Program (NVICP) handed down a decision in a mini-omnibus autism proceeding asking whether petitioners established “by preponderant evidence, a medical theory connecting a vaccine and [the test case child]’s injury.”
The decision is important in two ways. First, it reminds us that NVICP has consistently and repeatedly rejected claims that vaccines cause autism. Second, it explains in detail why a theory (please see Note 1 at the end of the article) claiming human DNA fragments in vaccines cause autism – a claim whose main proponent is Dr. Theresa Deisher – is unconvincing and not supported by the evidence.
The detailed, thorough decision shows that the main study from Dr. Deisher to support the theory – a study attempting to draw a temporal connection between change points where vaccines containing such DNA were introduced and rise in rates of autism – is fundamentally flawed. It then also shows that the petitioners’ proposed mechanisms of causation – how the DNA fragments are supposed to cause autism – are untenable.
The Mini-Omnibus Autism decision is 94-pages, and this summary will just touch on the main points. I urge readers to wade into the full decision if they want to understand more.
Mini Omnibus autism proceedings
In 2009-2010 the NVICP handed down six decisions in six test cases chosen by a steering committee representing over 5,000 cases in which claimants claimed vaccines caused their child’s autism. Two of the cases were appealed to the United States Court of Federal Claims and then to federal circuit courts. As the decision in this case explains, “All told, the 11 lengthy written rulings by the special masters, the judges of the U.S. Court of Federal Claims, and the panels of the U.S. Court of Appeals for the Federal Circuit unanimously rejected the petitioners’ claims, finding no persuasive evidence that either the MMR vaccine or thimerosal-containing vaccines could contribute in any way to the causation of autism.”
The other cases in the Omnibus proceeding had to decide what to do. Many were withdrawn or rejected. Twenty-three petitioners in this case came together to claim that their child’s autism was triggered by an adverse reaction to human DNA in an administered vaccine (In the test case, the alleged vaccine was MMR, but I have not looked at all the others and am not assuming they are all pointing to the same vaccines). The question in the case focused not on the specific child’s (referred to as V.J.M) medical history, but on the more general question of whether the causation theory is plausible. That question would be relevant to all the cases, and hence covers all 23.
To put this in legal term, to prove causation in cases before NVICP petitioners need to meet one of two sets of requirements. Petitioners can show they suffered a table injury – one of the enumerated harms listed in a table prepared by the Department of Health and Human Services, within a specific timeframe, in which case causation is presumed.
Or petitioners need to establish with a preponderance of evidence (50% or higher likelihood) that:
the vaccination brought about her injury by providing: (1) a medical theory causally connecting the vaccination and the injury; (2) a logical sequence of cause and effect showing that the vaccination was the reason for the injury; and (3) a showing of a proximate temporal relationship between vaccination and injury”
See Althen v. Sec’y of Health & Human Servs., 418 F.3d 1274, 1278 (2005).
This case focuses on the first part, providing a medical theory. The premise does not have to be established and supported by studies, but it does have to be plausible and supported by a preponderance of the evidence. Expert opinion can be the basis of that evidence if it meets certain requirements.
In other words, the theory is being held to a standard that is not too demanding, but not completely lax. It does not have to be scientifically accepted in the community or rock solid. It does, however, have to be plausible.
It wasn’t.
Vaccines and autism
In its review of the proceeding leading to the mini-omnibus, the Special Master reminded us that: “In none of the rulings since the test cases has a special master or judge found any merit in an allegation that any vaccine can cause autism.” (P. 9). In a footnote, the special master addressed two cases where compensation was given to a child with ASD. The special master said:
in neither of those cases did respondent concede, nor did a special master find, that there was any causation-in-fact connection between a vaccination and the vaccinee’s ASD. Instead, in both cases it was conceded or found that the vaccinee displayed the symptoms of a Table Injury within the Table time frame after vaccination and causation under the Act was presumed. (P. 9, footnote 15).
In other words, in both cases – one of which was the Hannah Poling case – compensation was given because causation was presumed based on a table injury – causation did not have to be actually shown.
This is a good reminder that, as was explained elsewhere, NVICP has never compensated a child on the theory that the child’s ASD was proven caused by vaccines.
Dr. Deisher’s claims
There are two parts to the discussion of Dr. Deisher’s autism-causation claim that was presented at the Mini-Omnibus autism hearing. First, Dr. Deisher tried to provide evidence that there is a correlation in time between introduction or mass use of vaccines made using cell lines and a jump in the rates of autism. Second, she provided theories to explain how DNA fragments can cause autism. The decision analyzed and rejected Dr. Deisher’s claim on both parts.
What are Cell Lines?
In essence, vaccines against viruses tell the immune system to make an immune response as if it was attacked by a virus, naturally infected, without the risks of an actual infection.
Viruses are parasites. They cannot grow without some living cell serving as a host. Whether a vaccine contains a weakened – attenuated – live virus, an inactivated – killed – virus, or part of a virus, a first step is growing a virus in something. That means at some point a virus needs to be grown in a live cell.
For complex reasons, some of the vaccines given to children use cell lines descended from abortions performed for other reasons – not for the purpose of creating vaccines – from the 1960s. The current cell lines are many generations removed from the initial cells, and the vaccines do not contain cells – the virus is removed from the cell. However, the vaccine can contain small amounts of DNA from the cells.
Although criticized, even strongly pro-life organizations and individuals do not generally see the use of cells as a reason not to vaccinate a child, even with the vaccines using the cell lines.
Dr. Deisher clearly has concerns about the use of cell lines (decision, p. 87-88). She has worked on the development of animal-based alternatives – by itself, a constructive way to address such a concern – but her hostility appears to have affected her research in other ways.
Dr. Deisher’s Change Point Study:
Dr. Deisher performed a change point study, trying to correlate changes in autism prevalence to the introduction of viral vaccines. She identifies as change points 1980.9, 1988.4, and 1996. Dr. Deisher collected data about autism and vaccines from the United States, Australia, U.K., and Denmark. She claimed the change points in autism prevalence corresponded to introduction/increased use of the human cell-line manufactured vaccines. Dr. Deisher’s study has been criticized before. You can find excellent analyses of Dr. Deisher’s studies here, here, here, and here.
But the decision, drawing on testimony from three serious experts (Dr. M. Danielle Fallin in autism epidemiology, Dr. Neal Halsey on vaccine science, and Dr. Dan Arkin in genetics), does a very thorough, thoughtful and respectful job of pointing to the flaws and problems in the study. I will only touch on the highlights here.
Dr. Deisher’s main argument is summarized in a graph included in the decision.
The special master explains why the study cannot support that vaccines cause autism:
First, as best explained by Dr. Fallin, the study design, even if perfectly executed, does not allow for an inference of causation. Second, there are concerns about the accuracy of the underlying data. Third, Dr. Deisher assumes that vaccines containing residual human DNA fragments, rather than other sociologic or environmental factors, are causing an increase in AD. Fourth, the statistical software used in the study may have been a poor fit for the data analysis.
The first point is that a study like Dr. Deisher’s that focuses on disease trends cannot be used to conclude causation, because population level links may be untrue at the individual level – the so-called ecological fallacy. The decision explains:
The ecological fallacy is Dr. Deisher’s conclusion “that vaccines using fetal cell DNA are a cause of autism.” Id. at 695. Like the example concluding that computer use causes asthma because computer use is seen in relationship to an increase in asthma, the assumption that there is a causal relationship between vaccines and autism is an ecological fallacy.
The ecological fallacy claim was further discussed by Orac before, and I will leave it aside for brevity.
The second problem is that the data Dr. Deisher used was collected from different sources at different points of time. It is generally hard to accurately measure autism prevalence and incidence data, and here, many factors make the data Dr. Deisher used especially suspect. The first is the change in diagnostic criteria over time – which have changed drastically, as reflected in changes in the DSM – the Diagnostic and Statistical Manual of Mental Disorders – over time, and affect rates of diagnoses. Dr. Deisher tried to account for that by calculating DSM printing dates to see if they correlate with changing rates of autism – and found that they do not.
But, as pointed out by Dr. Fallin, that’s a problematic approach, since the acceptance of the revisions was not immediate or uniform by practitioners – so assuming printing a new DSM was the same as the new criteria being broadly implemented and applied is incorrect.
Similarly, the databases used have changed their reporting practices, which would affect the rates of autism – the example given in the decision was the rates in Denmark, which initially counted only autism diagnoses in inpatients, and then switched to including outpatient data – leading to an increase of diagnoses. Dr. Fallion also pointed out that areas, where there is lack of access to services or a negative stigma associated with autism, may also see fewer diagnoses, that increase of awareness and increase in risk factors like parental age may lead to more diagnoses, and so forth.
Besides these additional factors, the assumptions on which a correlation between change points and the vaccines was claimed were problematic. For example, Dr. Deisher assumed that ACIP’s recommendation of a second dose of MMR was immediately implemented – Dr. Halsey said that because AAP disagreed and supported a second dose between 11-16, most children at the time got the vaccine in adolescence. (P. 36)
Similarly, although Dr. Deisher claimed a 1995 change point was correlated to use of the varicella vaccine, the vaccine was not recommended by ACIP for children until 1996, and uptake was slow. Further, there would not be an immediate change in autism diagnosis, because of slow uptake – there would be a delay.
In these circumstances, associating a 1995 change point with a vaccine licensed (but not recommended) in 1995 is simply impossible.
In addition to these problems, the experts for respondent in the Mini-Omnibus autism hearing pointed that Dr. Deisher likely misused the R program she used in her analysis. They criticized her lines because the way she drew them was not appropriate for data where most of the value is zero – as some of her data was – and some of the lines were mostly above rather than through the data. I don’t have the expertise to assess if that’s right. But the statistics experts, who presumably have the expertise, suggested that those problems made the lines a poor fit, and the change points unreliable.
Finally, Dr. Deisher ignored the many “well-designed observational studies” (P.39) that looked at whether MMR causes autism and found no link. After all, of all our vaccines, thanks to Andrew Wakefield’s misrepresentations, MMR was most studied for a link to autism – and none was found. The decision reviews several of those (pp. 39-41) and points out that they are a powerful body of evidence.
Dr. Deisher tries to counter that by claiming that MMR causes the “regressive form of autism”, and the studies did not focus on that. But that claim fails, too. First, the abundant evidence at the least cast doubt on the claim that there is any link between MMR and autism. Second, some of the studies – notably Taylor and Fombonne – did address regressive autism, and found no link.
The Special Master adopts the Special Master’s decision in Cedillo (pdf), that said:
the balance of evidence from those studies weighs against the petitioners’ causation theory. First, it is indeed an exceedingly slight point in the petitioners’ favor for them to claim that these many studies by different researchers in different countries have not completely ruled out the possibility of any merit to their causation claim. The larger point is that none of those many competent studies has yielded the slightest bit of evidence in the petitioners’ favor – and, of course, it is the petitioners’ burden to show that the MMR vaccine does likely cause autism, not the respondent’s burden to show that there is absolutely no possibility of a causal link.
Second, in [ ] view [of] the failure of so many studies to find any association between MMR vaccine and autism, while not completely ruling out a possible causal role with respect to a subset of autism, at least casts considerable doubt upon the proposition that the MMR vaccine ever plays a role in causing any kind of autism, including regressive autism. Id. at *89-90 (emphasis in original).
Causation theories
In addition to creating a correlation – as her study tried (and failed) to do – Dr. Deisher had to propose claims of how the DNA fragments in vaccines caused autism. Dr. Deisher’s theories fall into two main categories – “insertional mutagenesis”, which means that the fragments got into the cells, got into the genetic material, and somehow caused autism – and autoimmunity, which means that the child developed an immune response to the DNA that caused the child’s immune system to attack its own cells and caused autism.
The specific theories are complex, and the decision’s thorough debunking of them is, too. I’m not going to go over them in detail because this would make this already long post too long. I will, however, give some highlights.
A. Insertional mutagenesis
Basically, to support her theory, Dr. Deisher had to suggest a way to show that the fragments of DNA in vaccines get into cells and cause mutations in brain cells that they can affect cell function. She proposed four different mechanisms, lumped under the term “insertional mutagenesis”. The short version of these theories is that all four mechanisms she tried to use had serious holes, and the studies she brought in support did not support them.
For example, one claim she made was that a type of retrovirus sequence found in most humans – HERV-K – was also found in the DNA fragments in vaccines, and those fragments could enter genes and cause neuroinflammation. She draws on a study that found that retroviruses common to humans were elevated in children with autism – but Herv-K, in that study, was equal in the children with and without autism spectrum disorder (ASD). She also does not explain why a fragment of retrovirus – HERV-K – that’s already part of our genome would suddenly become an issue from vaccines: it’s not a new environmental factor, it’s already there. Studies have also not linked HERV-K to causing disease.
Another theory she proposed is called retrograde transport – a claim under which, in Dr. Deisher’s version, DNA fragments in vaccines are picked up by axons, part of a nerve cells that carry signals from one part of the body to another – transported by them to the cell bodies of the Central Nervous System and from them to the brain, where they cause mutations in human DNA. Dr. Deisher draws on studies about retrograde transport – but they don’t actually support her claims.
One study, Wang et al., used a special substance to allow DNA to get to the brain – suggesting, if anything, that DNA does not get to the brain by itself via injection. The injection there was also into the tongue muscle, “which is part of the rat’s hypoglossal nerve system” – on a direct pathway to the brain – and still, DNA alone did not get to the brain. Other studies showed that live viruses were transported to the nervous system – but the fact that live viruses can get to the brain via retrograde transport is not really evidence – that fragments of DNA can, or that injecting DNA into the muscle can transport across systems into the brain or cause ASD. (pp. 50-51)
A third claim called “microvesicle transport” is similar to “retrograde transport” in that DNA from a vaccine, which is then injected into the body, is taken up by blood and carried back to the nervous system. She bases her theory on an article from a scientist named Julian Bess, who only showed that microvesicles contain DNA which are the products of normal genetic processes in the human body. Nowhere is there any evidence that these microvesicles are capable of integrating non-self DNA into them or transporting them anywhere. She also does not seem to be able to keep retrograde transport and microvesicle transport clear or be able to distinguish the two.
The fourth theory – described as “most likely” by doctor Deisher – is the idea that “a hematopoietic stem cell in the peripheral circulation may take up a DNA fragment from a vaccine and insert it in the genome, causing a mutation in that cell.” Hematopoietic stem cells are stem cells that develop into blood cells, mostly located in the bone marrow, but a few circulating in parts of the bloodstream. Her argument is that these cells can pick up a part of the gene and insert it into brain cells, causing a mutation that leads to autism.
There are multiple problems with that argument. First, the articles she uses to support it do not. One of them, by McNeer et al, required creating a special tool to insert genes into hematopoietic cells – highlighting the fact that inserting DNA into cells is not simple or straightforward. If anything, it contradicts her apparent assertion that it can just happen. Her other example relates to trying to use gene therapy in “children with severe combined immunodeficiency (“SCID”), also known as “bubble boys.” – which tragically ended in the boys developing leukemia (p. 54-55). But here, too, inserting genes into the cells was hard – it required the use of retroviruses. Again, if anything, the difficulty in inserting genes goes against her claims (this point is further reinforced in pp. 57-58 of the decision).
Other problems with all of Dr. Deisher’s claims here are that mutations that happen postnatally will only affect the mutated cell and cells that derive from it – they could have a localized effect, like a tumor, but not cause diffuse brain effects like those seen in ASD (p. 62).
The experts on the respondents also pointed to numerous reasons to see the mutations that lead to autism as occurring prenatally – including the need to affect a large number of cells, and the type of cells affected. They supported that by several articles. (pp. 63-64).
B. Autoimmunity
The other mechanism Dr. Deisher proposed is that the DNA fragments lead some genetically susceptible children to create autoantibodies that attack the brain. She points to several studies by Mostafa et al. that found autoantibodies in children with autism. But that, too, runs into problems. First, Mostafa pointed out that the direction of causality is unclear – “whether autoimmunity to neuronal antigens is a consequence of the autoimmune diseases or actually initiates the process.” (pp. 69-70). Further, Neal Halsey explains that autoantibodies are found in people with or without autoimmune diseases – their presence is not enough for a clinical diagnosis.
In relation to varicella vaccine, one of her culprits, a study was done to check if the vaccine causes a harmful anti-DNA response, checking at six weeks and a year after – and the result was that no, it did not (pdf).
In addition, autoimmune diseases are not new, and with the extensive research done on autism, including on the immune system and autism, the claim that we would not identify it if autism was an autoimmune problem is unconvincing.
In short, this claim, too, is unsupported.
A final claim relied upon a study by Lindsay Walker which used diffusion tensor imaging (DTI) – an MRI technique – to look at brains of children with and without autism. The theory was that the varicella vaccine could cause an autoimmune response that attacks the brain, just as strep infection can cause certain neuropsychiatric disorders. Dr. Burkhard – a second expert who testified to this theory – claimed that the studies showed “showed dysfunctional myelitis in the brain, including areas of the brain responsible for executive functions.”
But, Walker herself warned against using her study to suggest that there is evidence demyelination. Postmortem studies of autistic brains do not show demyelination. If anything, they support the view that any effect on the brain was prenatal. (p. 73).
Conclusion on the Mini-Omnibus Autism decision
The Mini-Omnibus Autism decision addresses the claims made by petitioners very, very thoroughly and seriously, and shows how they are contradicted by existing studies, how they often rely on misusing or misrepresenting studies, and how they fail to provide evidence showing that DNA containing vaccines cause autism. After doing this work it found petitioners’ theories not plausible. They did not meet the relatively low standard set by the NVICP – a standard that, to reiterate, is less demanding than the one a civil court would set.
We are left, again, with problematic, implausible theories of autism causation that fly in the face of increasing data on the genetic origin of most autism cases, the evidence that it is prenatal, and the extensive data that shows no link to a vaccine.
The Mini-Omnibus Autism decision built a powerful case against this latest theory, reinforcing the point that looking at vaccines, when it comes to autism causation, is beating a dead horse.
Notes
- The term, theory, is used more in a legal sense in this article – lawyers and courts often use theory in a legal sense to imply a claim, assumption, hypothesis, or opinion. It should not be conflated with scientific theory, which represents a scientific principle that is supported by robust and overwhelming evidence, has withstood rigorous scrutiny, and embody scientific knowledge. Dr. Dreiser has not presented a “scientific theory.”
Professor Reiss writes extensively in law journals about the social and legal policies of vaccination. Additionally, Reiss is also member of the Parent Advisory Board of Voices for Vaccines, a parent-led organization that supports and advocates for on-time vaccination and the reduction of vaccine-preventable disease.
