May 23, 2008: The Government’s Response, continued
Direct Examination of Dr. Patricia Rodier
Dr. Rodier’s qualifications were reviewed on May 22, 2008. Mr. Johnson reviewed Dr. Rodier’s honors in graduate school and her NIH grants. She has received funding from NIH for ten years on the genotype and phenotype of autism and treatments for autism. The grants are for about $2.5 million per year. There are about 30-40 people at the M.D. or Ph.D. working on those grants.
Q: Have you ever testified before?
A: In writing, but not in court. I testified in the Canadian Omnibus case, similar to this one, and the Redfoot case last year.
Q: Is there a reason you didn’t testify in those trials?
A: Both of those cases were dismissed before trial.
Q: Can you briefly summarize the testimony you will give today?
A: There are many experts here, and I don’t want to overlap with other testimony. I think I’m the only person in the country who has worked both on autism and mercury poisoning. So I want to revisit the Bernard paper of 2001, which is the only reason that all of us are here today. That’s because it claimed that the symptoms of mercury poisoning are the same as autism, and then the second topic I want to talk about is when does autism begin; is it initiated prenatally or postnatally?
Q: You mentioned the Bernard article. Is this the article where the hypothesis that mercury causes autism began?
Q: As a scientist working in autism, do you have any criticisms of the Bernard paper?
A: I have many that are shared by scientists who work on autism and who work on mercury. There’s one published article replying to the Bernard article, and it’s by Nelson and Bowman. What they were criticizing was the characterization of symptoms of autism. The paper claims that nausea, vomiting, irritability, temper tantrums are symptoms of autism. They’re not diagnostic symptoms of autism; they’re things that happen to people with autism, just like to all of us. So they’re not useful diagnostic criteria. They also included symptoms like mental retardation, depression, abnormal gait – these things are not normal, but they occur in many conditions, not necessarily in autism. They’re not diagnostic symptoms; they can’t be used to compare autism with other disorders.
Q: As a scientist who has done work on mercury, do you have criticisms of the Bernard paper?
A: They look at vapor exposure to mercury; exposures to inorganic mercury and to pre- and post-natal exposures to methyl mercury, but few references to ethyl mercury, which is the only kind of mercury that’s relevant to their hypothesis, because thimerosal contains ethyl mercury.
Does the Bernard paper support a connection between mercury poisoning and autism?
It purports to, but it doesn’t. Let’s look at examples from the paper. [Dr. Rodier reviews certain points in the Bernard paper: depression; flat affect; mood swings; impaired face recognition. Dr. Rodier asserts there’s no overlap between autism and mercury poisoning. “These four things don’t go together in any way.” Discussion of echolalia; word retrieval problems; odd usages of words in Aspberger’s syndrome. Then discussion of vision, joint attention – Dr. Rodier asserts that the comparisons at Table 2 in the Bernard article are invalid.]
Q: Why don’t the authors of the Bernard article talk about the symptoms of ethyl mercury poisoning?
A: Good question; ethyl mercury poisoning is not as good a story. Zhang studied 41 people who had eaten rice tainted with ethyl mercury; they knew the dose. They examined them; they documented their symptoms. They counted how many people had each symptom. The effects were mild to death. They had a range of symptoms to examine. They found the three most common symptoms were muscle weakness, loss of appetite and dizziness. Those don’t sound much like autism. The next ten symptoms were nausea, abdominal pain, diarrhea, fever, numbness of the extremities…. Again, I think you can that see none of these sound anything like the diagnostic criteria used for autism, so you can see there’s no correspondence between ethyl mercury poisoning and autism.
Q: Petitioners claim that low levels of mercury persist in the brain causing oxidative stress or neuroinflammation leading to the symptoms of autism. Does that hypothesis make sense to you?
A: No it doesn’t. Scientists try to disprove a hypothesis. When this paper was written, and today, one piece of evidence disproves the hypothesis. In known cases of ethyl mercury poisoning, like the tainted pork case in New Mexico, people did get accumulations of inorganic mercury in the brain, shown on autopsy, but the neurological symptoms disappeared completely; the effects went away.
Q: Let’s discuss when autism begins. You identify five environmental risk factors.
A: These come from population studies: rubella; thalidomide; valproic acid, a seizure medication; ethanol; and mesoprastol. I’ve listed the times when exposure has to occur for the symptoms to result in autism. [All exposures are in early pregnancy.] It’s more likely that as we discover other environmental causes that they will be in this time frame.
Q: Do you know what tributelene is? Petitioners have raised this; why didn’t you include it on your list of risk factors?
A: The Connors study is a genetic study; it’s not a population study. They looked at twin pairs that had been exposed to tributelene. Only when they got to a very, very small subgroup of male twins with no other affected siblings did they find a significant correlation. Even there, it’s hard to ascertain the effect because tributelene is only given to avert premature births, so we don’t know if the autism was from the tributelene or would have been associated with the prematurity. [Discussion of just published article of low birth weight children and higher rate of autism; article not filed by either side] [Discussion of rat study where animals were given tributelene right after birth.]
Q: Have you studied brain samples from autistic individuals?
A: Yes. [Dr. Rodier points to slides showing brain autopsy of autistic child. She shows lack of nuclei in the autistic brain in facial expression area; she suggests that these nuclei form very early in gestation; also discusses an example regarding the Purkinje cells showing autism starts early in gestation. Discussion of Dr. Rodier’s study of dismorphologies in autism; these dismorphologies occur prenatally, including from environmental factors.]
Q: Do you think this evidence you’ve discussed is inconsistent with regressive autism?
A: No. Regression can follow an early injury.
Q: Do you believe that TCVs cause or contribute to ASDs?
A: I do not. There is no similarity between the symptoms of mercury poisoning and autism and there’s nothing like a similarity between ethyl mercury poisoning and autism. It’s my belief – from what’s known that risk factors, histology and dismorphology – suggest that autism arises in the embryo in the first trimester of pregnancy, and there’s no evidence that it arises postnatally.
Q: And do you hold your opinion to a reasonable degree of scientific certainty?
Cross Examination of Dr. Patricia Rodier [Mr. Powers]
Q: Did you review the Petitioners’ reports on general causation?
A: Yes, I did.
Q: Did you review Dr. Deth’s report?
A: Yes, but I’m not a biochemist; it’s outside of my area.
Q: Did you review Dr. Greenland’s report?
A: Yes, but it’s outside of my area.
Q: Dr. Aposhian’s report?
A: I did.
Q: Do you know that the Bernard article was not referenced there?
A: I don’t remember.
Q: Did you have a chance to review Dr. Kinsbourne’s report after you submitted your report?
Q: The Kinsbourne report didn’t discuss the Bernard the article at all, is that correct?
A: I don’t think so.
Q: So, is it correct that none of Petitioners’ expert reports discuss the Bernard paper?
A: I don’t remember, but, if you say so, yes.
Q: It seems that you’ve looked at minor physical abnormalities (MPAs) and autism?
Q: You’ve also done brain pathology work?
Q: How many brains have you autopsied leading to peer-reviewed publications?
A: Just one.
Q: One brain?
Q: In the literature that you’ve reviewed, would it include the same ones as Dr. Kemper referred to yesterday?
Q: That work involved 23 brains?
Q: So the brain pathology work that you’re relying on is 23 brains in those studies and the one you worked on.
A: I’m sure Dr. Kempner was including the one in his 23.
Q: In those neuropathology studies, there was no correlation made between the symptoms and the person from whom the brain was taken, is that correct?
A: I don’t recall very extensive discussion of symptoms, but at that time, all of the brains met the criteria for autism.
Q: So you can’t correlate the brain with any particular mix of symptoms?
A: You certainly can’t account for all of the symptoms with the neuropathology. We did see some correlation, however. [discussion of one case]
Q: So in this one symptom, you had a correlation. But in general, there is no correlation?
A: In general, yes. [discussion of Dr. Rodier’s eye blink conditioning study]
Q: Did any of your work on neuropathologies correlate with diagnostic behavioral criteria?
A: No, because none of the symptoms used in the diagnosis can be reduced to any one section of the brain.
Q: Is there anything in neuropathological findings that correlates with specific symptoms?
Q: [Questions about Dr. Rodier’s expert report on environmental factors] You say that the known environmental risk factors occur in the first trimester. What attributes are there to “be known”?
A: It’s been shown in population studies.
[Discussion of postnatal case studies; Dr. Rodier’s genetic studies on certain chromosomes associated with autism; her studies on dismorphology and autism]
Q: Do you disagree that there may be environmental factors that come into autism?
Q: Do you agree that it is biologically possible and plausible that there might be postnatal factors that result in autistic symptoms?
A: It’s a very outside possibility that a late injury could result in the same symptoms as an early injury. [Discussion of NIH recommendation to study environmental factors and autism.]
Q: What percentage of children with regressive autism have dismorphic features?
A: I’m not aware that anyone has studied dismorphologies in regressive vs. non-regressive cases.
Q: Among cases that are non-regressive, what percentage have dismorphologies?
A: In our sample, it runs around 50%. In the cases in Rome, collected by Tony Persico, he says 52% in his sample.
Q: So it sounds like about ½, and that’s in the non-regressive cases. And in the regressive cases, you’re not aware of any evidence associating regressive autism and dismorphologies?
A: His data and ours would have included some cases of regressive autism.
Q: Those studies made no distinction between regressive and non-regressive. This is an inquiry that has yet to be made?
A: That’s right.
Q: Based on your findings on one brain, did your results make sense biologically?
A: Yes. We had only one case, but we did serial studies on that brain. It wasn’t that expensive to do serial samples. In Dr. Kemper’s study, their sections are incredibly thick. In Bailey’s case, they also had very thick samples; there are a lot of parts of the brain they couldn’t study. [further discussion of examples of consistence and inconsistency between serial studies]
Q: Why did you pick the environmental factors you did -- because there is good epidemiological evidence?
Q: What’s more likely as a cause of autism -- an early prenatal injury or a postnatal environmental cause?
A: Early injuries produce a cascade of further injuries in the nervous system; I would favor the idea that early injuries leading to autism.
Q: We’ve read reports of people with post-natal exposures to certain viruses or other factors; herpes encephalitis, for example. Given your testimony that the injury probably occurs in early prenatal development, how do you square that with evidence of postnatal events can cause similar symptoms?
A: In those rare cases, like malaria or herpes, what you get is tremendous brain damage.
if you damage the brain enough, you’ll cause symptoms of autism, although it wouldn’t technically meet the autism diagnostic criteria.
Q: So you’re saying a post-natal insult can cause autistic symptoms?
A: Yes, but the symptoms of autism can occur in persons who have no other symptoms at all. Many cases don’t look like overwhelming brain damage. [Example of boy with 150 IQ with no other symptoms except autism] That’s the thing we’re trying to understand.
Q: Let’s say a child suffers a prenatal insult that takes away the ability to see. Does that happen?
Q: Then a child suffers some sort of brain inflammation or brain injury. These are very different causes, but similar results.
A: I understand the situation you describe; vision can be lost at any age. What’s different from the vision analogy, most people think the big part of the problem in autism is that the connectivity of the brain is not right. It’s not that there are holes; the connections didn’t form properly. That basic process is going on like crazy in the embryo and in the first few months in utero. The basic connections in the brain are all present soon after birth. To disrupt those pathways would have to take place early. I have trouble imagining is that you could then go in later with some kind of global injury that would give you the same misconnected brain that you get early.
Q: If you took these two children in my example, neither can see, what you’d see in the brains would be very different? Can you get similar symptoms to early injuries from a postnatal insult that don’t relate to those findings?
A: I don’t think we know the answer to that. It would depend on whether the problems were those of connectivity. For example, you could be blind for lots of different reasons – but they could have the same symptoms but through different mechanisms.
Q: But if you looked at the brain, you would see different things?
Q: So you could develop autistic symptoms postnatally through other mechanisms than those you’ve described?
Direct Examination of Dr. Steven Goodman, M.D. M.H.S., Ph.D. [Ms. Ricciardella]
Dr. Goodman is on the faculty at Johns Hopkins; works in biostatistics, oncology and epidemiology. He serves on editorial boards; advisory committees and lectures on inference and evidence. He’s engaged in research and directs a biostatistics department in oncology. The role is to collaborate in the cancer center and throughout the medical school. He has published over 100 peer-reviewed articles. The primary area of research is cancer. He has done research on peer review. He wrote the chapter in the 2004 Surgeon General’s report on smoking. He is the Editor in Chief of a journal called Clinical Trials. He has testified twice in court. Once in a medical malpractice case; once in the Phen-Phen case; he advised on whether the settlement reached was fair and reasonably based on the evidence.
Q: Have you every testified on the epidemiology of autism?
Q: Your CV shows that you have served on several committees of the IOM. Please tell about the IOM.
A: The IOM is an independent body chartered by Congress. It’s our job to read all the reports; the body of the report is written by the staff, but it’s guided by the members. All the language is reviewed by all of the members of the panel.
Q: Are IOM reports peer-reviewed?
A: Yes. Then it’s reviewed by the head of the IOM.
Q: You were on the immunization safety review committee?
A: Yes. It was formed by Congress and the CDC about various hypotheses that the evidence underlying these hypotheses be adjudicated by science. We first met in 2001.
Q: How many members were on the committee?
A: 13; I think 13 medical institutions were represented. The Committees started meeting in 2001, and they issued a series of reports. The first report was on MMR and vaccines. The second was on thimerosal and developmental disorders. Then polio and SV-40 and cancer. There were 8 reports. The final one again revisited thimerosal and autism.
Q: Do you recall the specialties that were represented in the Committee?
A: Neurologists, neonatologists, immunologists, epidemiologists, biostatisticians, public health experts and vaccine biologists.
Q: There were no toxicologists?
A: There were not. There may have been toxicologists among the reviewers.
Q: Do you think it would have affected your conclusions if you’d had one?
Q: Have you been in other IOM Committees?
A: Yes. The Committees on Agent Orange and post-traumatic stress disorder in soldiers returning from Iraq.
Q: Regarding the 2001 Report, what conclusions did this group make regarding thimerosal and vaccines?
A: Our conclusion was that the evidence was inadequate. The reason was that there were at that time no epidemiological studies. It recommended further studies.
Q: In the first report it said the mechanism was “biologically plausible.” This was taken out in the 2004 Report. Can you explain why it was taken out?
A: The change in language is described in detail in the 2004 report. The idea that mercury could cause neurological damage was plausible. It didn’t violate any known biologic rules. Then the Committee saw that it was too vague and was interpreted as if there might be a link. So in the 2004 report, it says the idea is “theoretical only.” It’s not impossible.
Q: Why did the Committee meet again in 2004?
A: The main reason was that there was now a moderate amount of epidemiological evidence to consider since 2001. The IOM wanted us to weigh in on that. The Committee itself did not decide what we were to address.
Q: What was the clinical outcome that was studied by the 2004 Committee?
A: Autism and ASDs.
Q: What kinds of evidence was presented?
A: We saw a range of evidence; the epidemiological evidence; we also saw laboratory, clinical, animal studies that related to the hypothesis that thimerosal caused ASDs.
Q: Was the public invited to comment?
A: Yes; we had public session.
Q: What conclusions were available to the Committee?
A: Proven; the evidence favors the conclusion; inadequate evidence to make a conclusion; there’s no evidence; favors rejection of a causal relationship.
Q: What conclusion did the IOM Committee reach?
A: We favored rejection of this hypothesis.
Q: Was it unanimous?
Q: When is this “favors rejection” conclusion used?
A: It’s used when all the evidence points away from a causal relationship and there’s no mechanistic evidence that would contravene it. It absolutely doesn’t mean that it rules out the possibility of a relationship. It’s just a conclusion that says that the evidence weighs in on one side. We thought it was very important because we were talking to the public and the public health agencies. We thought it was important to address parents; I had a child at this point. It had a big emotional impact; we were aware that we had to speak – there were high stakes for the parents as well as the public health community.
Q: Does the IOM usually come to this conclusion?
A: No, they generally don’t. They don’t usually say “favors rejection.”
Q: The IOM said no additional resources should be allocated to this line of research of TCVs and autism; is that an accurate characterization?
A: No, not at all. We felt that we didn’t understand the biology and risk factors very well. It was hard to target any epidemiological groups; to continually go back over looking at the same data was problematic. The experiments were very expensive. We said something like, the money should be funneled to the most promising areas of research. It was a question of prioritization.
Q: You’re an epidemiologist?
Q: What is it?
A: Determinance of patterns of disease and the risk factors for those diseases.
Q: It’s the only science that looks at an exposure and at outcomes in humans. Epidemiological type designs under gird almost all science in medicine. Epidemiological evidence can also be said to include experimental science.
Q: Is any epidemiological study perfect?
A: Well, epidemiology is observational study. So, with smoking, you see people who smoke or don’t and see the differences in outcome. The challenge is to figure out if the difference you see in the outcome is from the exposure or something about the people themselves. So, if all short people smoked and all tall people didn’t, we couldn’t easily figure out what the impact of smoking was. Every study can be subject to criticism.
In the end, you have to do different studies in different ways, measuring different populations. What is believed in the field is that the kind of bias that any one study has is not reproduced in all the others, so we try to find the amalgam of evidence that makes a different outcome less likely.
Statistics are very important, but it’s the marriage of statistics with an understanding of the underlying biology. You wouldn’t know where to start without a sense of the underlying biology. It also helps when you get unexpected findings; it helps to figure out if the results are spurious or meaningless. Biology is at every step.
Q: What is required for a causal determination?
A: You want to see a relationship beyond the world of chance where, at a minimum, it doesn’t violate any biological or physical rules and at a maximum has a coherent relationship behind it. In smoking, the relationship was so strong and so compelling, it was hard to resist finding causation. In most cases where you see results that are weaker, it still has to be beyond the world of chance. Still, it has to be beyond the play of chance and you have to rule out alternative explanations by doing many different types of studies. It’s only epidemiology where we really address the central question, is a human being exposed to thimerosal at higher risk of developing autism? Any other science is going to look at pieces.
Q: How many epidemiological studies did the IOM consider?
Q: How many epidemiological studies have there been since the IOM 2004 report?
A: I think it’s 4.
Q: I’d like to turn to Dr. Greenland’s report. Did you listen to his testimony?
Q: I’ll refer first to his written report and then his testimony.
A: The existing epidemiological studies don’t rule out that there could be a subgroup. The epidemiological evidence dilutes the effects on that subgroup. It comes down to the view that it’s possible. I don’t disagree with that. “I could have written it myself.”
Q: Is the subgroup he’s referring to “clearly regressive autism”?
Q: Would you please explain mathematical bounds.
A: The bounds that he calculated are the approximate limits of how high a risk might be in that subgroup that’s still compatible.
Q: Is that what Dr. Greenland is doing in this case?
A: Yes. There’s always going to be some residual uncertainty.
It’s always going to be the case that you can include a 0 effect and we can fit a high-risk, small group. It’s mathematically possible. It’s correct. It’s algebra. Where Dr. Greenland doesn’t go, and the endpoint of my report, is to go from the possible to the probable. That’s where the IOM Committee thought it had to go. He doesn’t say it’s probable; he opined that it’s possible.
Q: You talk about the confidence intervals that are two-sided.
A: So it’s the imprecision; they give the range of true relationships consistent with the data. You could say there was a protective effect. So the math itself doesn’t differentiate between a strongly protective effect or risk. Most of the studies show a protective effect. Do I necessarily believe that? No. I don’t think there’s any biological evidence. So, the reason we don’t say it’s protective is because we don’t think it has strong biological plausibility, at least we don’t think it has an excess risk.
Q: Taken together, what does this suggest about an autism epidemic?
A: Again, Dr. Greenland and I are in accord; any thimerosal effect would be nil or small; it would basically rule out large increases. What they don’t rule out are small subgroups. You can’t have it both ways. They make it almost impossible to implicate thimerosal in wide-scale autism.
He calculated these bounds, saying let’s imagine this just raises the rate of regressive autism, and let’s say that’s 10%. But it assumes that it had all of its effect for “clearly regressive autism” and 0 effect for the other 90% of the population. This calculation absolutely requires that most or all of the element of risk is restricted just to this one subtype and none of this excess. So he’s positing a dramatically different pathway. That’s the basis for his calculation.
As soon as you start to allow a little bit of extra risk for everybody else, that distinction requires some kind of biological or mechanical explanation. We’d have to have a group that has a fundamentally different biology. Whether that’s possible, yes; whether it’s probable, there’s no evidence presented. This was acknowledged in the IOM report.
Q: He used the analogy of cancer. Is it a proper analogy?
A: Yes; it makes my point. If you asked me to distinguish two types of leukemia, yes, I can show you – you can look at a slide. They have different biological features; they have different treatments. If you ask me the difference between two types of lung cancer, they are biologically different. That’s what makes different risk factors completely plausible. It’s not the same in the autism world. It may have to do with things we can’t imagine today. Is there any evidence that these two forms – classic autism and clearly regressive autism are causally different? Not today.
Q: What is meta-analysis?
A: It has two parts; a quality assessment of the studies themselves and then an adding up of the studies and the precision of the combined estimate is almost always more precise than either alone.
Q: What would a meta analysis likely show?
A: They would show that the upper confidence limits were probably a lot closer to 1; the reason the IOM Committee didn’t do that is because there can always be questions about how comparable the studies are. What you can say qualitatively is that combining the limits would make them more precise and closer to 1 or a 0 effect. We didn’t do that at the IOM because you’d get arguments about where that limit is.
Q: Dr. Greenland did not submit a rebuttal report, but he did criticize your report. Do you have any comments on that criticism?
A: As he said, these subtypes were not measured in these studies. One cannot rule out that possibility, but what is the evidence that makes it likely in the first place? He didn’t offer any evidence. So, it still remains mathematically possible, whether it’s likely or probable, that’s another question upon which he did not opine.
Q: During his testimony last week, he also took issue with your use astrological signs and called it “rhetorical nonsense.”
A: He’s right that it’s about why you have to have biological plausibility to make something credible. They had a study in 11 astrologic signs and all the effects were under one sign. You wouldn’t believe it and think it’s implausible because there’s no biological underpinning.
Here we don’t see anything. The contention is, maybe you’d see something if you looked at the subgroup – we don’t even have that subgroup to look at; that’s acknowledged. The point was that the math doesn’t tell the story all by itself. In this case, we don’t have a pattern we have to explain away. The point of the example was to say it’s very important to have some biologic, mechanistic rationale to support a hypothesis to begin exploring it. Then you look for information to make it probable.
Q: Dr. Greenland accused you of invoking “fictional scientific principles” that you were presenting absence of evidence as if it were evidence of absence. Is that what you’re doing here?
A: I am completely open to having this being empirically proven. Maybe the scientific community will take this seriously simply because there is so much concern about it. But it’s a curious statement when one is positing no evidence for it. The fact is we do have evidence from overall population that shows no effect. That is what we do with clinical trials. It is always possible for someone to come in and say it doesn’t apply to this kind of person. Very, very frequently in medicine we find this claim of subgroup effects doesn’t hold up. It doesn’t rule out the possibility of different effects among parts of the population, but I don’t think it’s fair to ask for a rebuttal when there’s no evidence for it. Maybe this study will be done. But I don’t think the hypothesis remains likely or probable. It remains as a hypothetical. I didn’t say that it couldn’t be true.
Q: Does Dr. Greenland state that TCVs cause “clearly regressive autism”?
Q: Has he established that there is a different cause?
Q: Did he offer any evidence that TCVs elevate the risk of “clearly regressive autism”?
Q: Why are you testifying in this litigation here today?
A: Some people ask why I served in the IOM Committee too. This is clearly an area of a tremendous concern and amount of heat. If scientists don’t speak out, looking at it objectively, actually seeing the evidence, I feel like they’re not serving the majority who want to do the best thing for their child and for the parents of autistic children to feel that they didn’t cause their child’s autism. I feel it would be very easy to duck these kinds of activities. But it wouldn’t serve the public health and the really very sincere concern of parents on all sides to not say honestly what we think is the most likely message from the evidence for what we see today. And to point the way for further research that might elucidate the causes and cures or treatments for this very, very difficult disease.
Cross Examination of Dr. Goodman [Mr. Williams]
Q: You mentioned that you’re editor in chief of a journal about clinical trials in general. How long have you been the editor in chief?
A: Since 2004. I was appointed in 2003.
Q: Who is in the Society for Clinical Trials?
A: It involves folks from government, academia.
Q: Do you ever have to go to meetings for the Journal? Does the Journal pay for your travel?
Q: Are there corporate sponsors?
A: It’s not a corporate society by any means. There are academic and corporate sponsors for the Society’s meetings.
Q: I just want to point out that two of the corporate sponsors are major manufacturers for vaccines.
A: I did know there are corporate sponsors for the meetings.
Q: Do you ever do consultancies for corporate sponsors on vaccines?
A: I may have in the past on the design of clinical trials; I might have visited a company 10 years ago, but I honestly don’t remember the details.
Q: In the IOM 2004 report, it stated that the representatives on the Committee had no conflicts of interest. You don’t consider having vaccine manufacturers as sponsors for your journal a conflict of interest?
A: They support the meeting of the Society of Clinical Trials; it has no bearing on the actions of the Society. The answer is no.
Q: Have you looked at Dr. Kinsbourne’s report?
A: Yes, but I’m not expert in that area.
Q: I take it you haven’t reviewed the infant monkey studies?
A: I’ve read them, but I’m not an expert in this area.
Q: Is it still your opinion today that the biological plausibility of TCVs causing autism in some kids is as silly as astrology?
A: I never said it was silly, but in order to elevate it from the possible to the probable, you need more empirical evidence and a more accepted theory than we have today. I rely on the scientific community to assess Dr. Kinsbourne’s hypothesis. If I was part of the scientific community evaluating the paper, I would look at his published papers and the response of the community to it.
Q: You realize that in this vaccine court, it’s very rare that these cases get written up and published in the peer reviewed literature don’t you?
A: Right, but I assume that the underlying principles on which you’re making a judgment apply in the scientific literature.
Q: In the 4 or 5 epidemiological studies the IOM had, most were ecological, right?
A: 2 or 3 were ecological, and 2 or 3 were controlled studies.
Q: And in your report you say that the studies give “strong” support to reject the TCV – autism hypothesis. You’re using ecological studies as part of that strong support?
A: Yes. Reasonable people could disagree about whether it’s strong or moderate.
Ecological studies are not as strong as the control studies that we looked at, however, where the exposure from the vaccine goes to 0 and rates still seem to continue completely unconnected with that – those are subject to less concern. The general point you make is right – ecological studies are weaker studies than the controlled studies. It is extremely interesting that the autism rates seem to go up in a context where thimerosal exposure has gone to 0 in some countries and here is very low.
Q: And you don’t know about “clearly regressive autism” in other countries or in California?
A: Right. You can measure all sorts of things. You have to restrict the things you measure. The reason they didn’t measure it was that this hypothesis wasn’t out there. It’s not a complete accident that that information is not there. It would be very hard methodologically to make that break-out determination. You have to have a lot of clinical evidence; it would be an extraordinarily difficult study to do.
Q: Do you have an opinion about whether the rate increases in California are real or an artifact?
A: I would say Dr. Fombonne is far more expert. If rates are going up at all, I think it is an unsettled scientific question. I believe strongly that it’s not gone up astronomically.
Q: Did you review the NIEHS 2006 studies about how vaccines could be studied in the VSD?
A: No, I didn’t.
Q: NIEHS convened a group of experts on autism to discuss whether to do studies using the Vaccine Safety Datalink. They recommended that two specifically be done; you weren’t aware of this?
A: No; this is not my field. I followed this through 2004; I follow generally now; I don’t follow all the details.
Q: You see it’s signed by the Director of the NIH? I want to turn to the list of experts convened. Do you know these experts [He knows some; he respects them as epidemiologists.] They recommend a study of high risk individuals; siblings of people with ASDs; and an expansion of the VSD study published by Verstraten with several additional years of data. Do you agree that those studies would also be a good idea?
A: Probably; but I don’t have to deal with budgets and competing priorities. IOM itself recommended further research in high risk populations.
Q: Unfortunately, the Bush Administration has not chosen to fund these studies. The Petitioners’ Steering Committee did try to get access to VSD information through a motion. Wouldn’t you agree that one of the reasons we don’t have epidemiological studies is because the government has blocked us from getting those studies?
A: I have no basis on which to render an opinion on that.
Mr. Matanoski: I object to that characterization of what the government has done.
A: In theory, going back to the NIEHS recommendations, if we had a good study of those issues that might be good, but I have no access to the real world considerations that go into this.
Q: You postulated that the group affected by TCVs would be very small.
A: I followed what Dr. Fombonne and Dr. Greenland said.
Q: Roughly 4 million kids were born every year from 1992-2001, the years of biggest thimerosal exposure, so roughly 40 million children. If the true rate were 1 in 150, then if TCVs were causing 10% of the total, which is the figure you’ve used, the relevant number is the proportion to the total. If it was 10%, then it would be 1 in 1500. Let’s say that TCVs only account for one third of that regressive group, so our rate would be 1 in 4500 – you come up with 9,000 kids. You still think that’s a small group?
A: Small is in proportion to the whole population. I’ve never made any statement about the number affected. Autism is a very big problem in this country. 1 in 150 is a very high rate for a problem of this magnitude. Issues related to autism are important for this reason.
Q: And issues related to etiology are important?
Special Master Hastings:
Q: You mentioned that you read the report of Dr. Kinsbourne.
A: I read it, but I have not studied it.
Q: Dr. Kinsbourne went a little farther than Dr. Greenland. Dr. Kinsbourne said that the epidemiological studies are totally irrelevant to the Petitioners’ theory of regressive autism. I want to make sure I understand what you said today. You’re saying that the epidemiological studies are relevant because they don’t mathematically rule out the possibility of a regressive autism subgroup, but you’re saying because it makes it seem very unlikely that there is an association with such a subgroup because we don’t have any reason biologically to assume that there would be a difference.
A: That’s pretty much exactly it. The information on overall relationship is relevant to all children with autism until one can make and show empirically that one subgroup is uniquely biologically different with respect to that causal factor. It’s only irrelevant so long as he can empirically demonstrate that this is a unique biological entity that’s uniquely susceptible to thimerosal where all other children with autism are not. You described it pretty much exactly.
Q: It wouldn’t have to be just thimerosal as a postnatal insult? There could be a susceptible subgroup that gets postnatal insults and gets the symptoms of regressive autism?
A: I can’t disprove it. That’s one of many possibilities.