Senate Rural and Regional Affairs and Transport Legislation Committee [Part 2], , 2:43:12 to 3:26:01
The APVMA defends its methodology for weighting scientific evidence regarding the link between Paraquat and Parkinson's disease. Senators challenge the exclusion of certain systematic reviews and the influence of industry-sponsored research.
Topics: Health, Integrity, Agricultural research and levies
Speakers: Scott Hansen, Senator David Pocock, Unknown Speaker, Speaker 18, Speaker 5
>> Um thank you for your time this evening. U Mr. Hansen, I wanted to talk um Paracquat and looking at your decision, >> I wanted to walk through the APMA's analysis of the published literature as it relates to to Parakquot. Um, I wanted to start with um Brent and Schaefer uh 2011. Um, and I've I've got a copy to circulate um if it's helpful for for people. >> The APVMA, if
>> it's okay, I might just call up one of my colleagues >> if we're going to go by paper to that
states Brentton Schaefer 2011 examined the entire body of published reports of clinical outcomes of survivors of parakquot poisoning and found no connection between highdosese parakquot exposure in humans and the development of Parkinson's disease. However, what the APM may failed to state is that Dr. Brent at the time he authored that review was a consultant to Sententa who manufactured parakquot. Um and they you also failed to to state that the long-term survivors included any person who lived beyond 30 days Parkinson's has a latency of years to decades and this study looked at people just 30 days after poisoning. Um in fact my team went and found a letter to the editor that Dr. Brent had written about his study where he acknowledges himself that no conclusion could be made about chronic lowd dose exposure to parakquot and its association of Parkinson's as a a science-based agency. What would you say the limitations of this study >> because there weren't any mentioned in your regulatory documents?
>> No. And and so just while um my colleagues Mr. Dellal uh pulls that individual paper up because as you'd be aware we literally looked at hundreds of studies. Um and in our um additional neurotoxicity report that we produced um to go along with the uh reg final regulatory decision, we explained the approach we took with regards to the waiting of evidence and how we treated and viewed evidence in the we we consider all studies and what we do is we apply a particular waiting as to what does the study show us? Does it add anything new to the story? Does it does it shift the regulatory decision that was found in 2016 from the Department of Health's assessment for us on neurotoxicity? um concerns and if so how does it shift it and what do we need to do with that regard so on that case and I think we talked about this but I was talking in riddles in the last time we spoke because it was about two weeks before we were due to make our decision so I was um um circumspect about it >> but there's a whole volume of research a large part of it is drawn off the same data set of the US agriculture health which was set up by the US National Institute of Health and followed farmers and spouses in Iowa and North Like that's probably the biggest data set that exists that actually traces and tracks activity and behavior of farmers and their families and then their health
outcomes. I've got a bunch of I I'll get to some of those studies, but I'm interested to this one because reading your um everything you put out, you um seemed happy to point out the shortcomings of studies that showed an association of Parkinson's, you said, "Oh, well, these things don't." But then this one, which is 30 days, um you used as a study that showed that there was no >> but I wanted to turn to your exclusion of systematic reviews and meta analyses.
>> Yeah.
>> Um, and the APV's position is and and I'll quote um cited references presenting reviews, opinions, editorials, commentary, or clinical opinions or advice were excluded from detailed review. Um but as so many scientists have told me um over recent months, systematic reviews and meta analyses occupy the highest tier of evidence in evidence-based medicine and toxicology. And some of the systematic reviews you excluded on the basis that they were reviews such as Pizolei and Sarata 2013 uh pulled and analyzed 104 of these studies and concluded uh two times risk for Parkinson's associated with exposure. The thing I really take exception to is that you didn't even apply your own rules uniformly because the review titled does parakquot cause Parkinson's disease a review of reviews very clearly a review um was included and so is weed 2024 titled a systematic assessment of recent epidemiological evidence. Neither of these contain original primary data, and yet both are cited in your executive summary as supporting the APVMA's conclusions. So why are you excluding reviews that find risks of Parkinson's, but you still site reviews that support your conclusion, but you say up the top that you're not going to look at reviews? Uh
so reviews and meta analysis that don't present any new data that predate previous regulatory assessments that we undertook >> were generally excluded from the
analysis. And so why did you site them?
>> The ones that were post
>> why did you cite them as supporting your decision if you
>> but ones that were post >> were included because I
>> sure
>> answer question
>> if it's the answer to the question very happy to
>> yeah but I think he was he's answering your questions I think you should just allow him to answer before you ask
another question all the time you need Senator just
>> okay I'll take that if I've got plenty
about if we need to drop two seconds. >> We we'll try let's not talk over each other. Everyone gets a chance. Thank you. >> Uh sorry, Senator, were you asking the question or was Mr. Hansen answering?
>> Oh, I was in the middle of answering.
>> Mr. Hansen, go for it.
>> Yeah. So >> where the scientific reviews or the meta analysis actually introduce new data or bring something new to the table that wasn't previously considered in the previous assessments then they were referenced and included in whether it was positive or negative. So those ones that actually brought new information were included in that assessment, but meta analysis from the past that had previously been considered and/or brought no new information to the table
>> But that wasn't your approach. You say um cited references presenting reviews, opinions, editorials, commentary or clinical opinions or advice were excluded from detailed review. But then you cite reviews in your executive summary to support your decision. So
>> well we cite a lot of them.
>> But you've told us you're not looking at reviews but then you cited reviews to
decision. >> Where we have where scientific reviews been quoted in the report the citation has been included in the reference list for people to refer to it. that the approach we've used is completely consistent with international regulatory practice and with the inclusion and exclusion criteria that was used also by the US EPA in their reports. >> so so can we go back to the Brent Schaefer question that you asked that we didn't get around to answering. So um I might hand over to my colleague in a minute but the I mean so that study did only look at parapot poisoning survivors sustaining a single event highdosese oral exposure but it is one of the few studies that exists in which there was actually known and not negligible exposure that was documented and known to exist. So when we're looking at safety, we obviously look at acute toxicity as well as neurotoxicity. Um which is why that sort of paper gets a reference. Um but I might hand over and see whether Mr. Teller has anything in addition he wants to
>> but you sorry but can I just go come >> um yes, if you were looking at acute toxicity, I'd accept that. But you say found no connection between highdosese paracquot exposure in humans and the development of Parkinson's disease. This was a study looking at exposure after 30
Yeah. So irrelevant as it did analyze hundreds of subjects with a known very high exposure to
>> 30 days though. >> Parkinson's takes years.
>> Yeah. So
>> and old mate then then says hey you definitely can't take that out of this study but the APVMA has taken that out
of this study. But if paracquot was able to act by any mechanism comparable to breaking that bloodb brain barrier that is the the cause of the the hypothesis about what it breaks and gets through induced to induce Parkinson's disease like it might be expected to be observed in survi subjects surviving an acute highdose poisoning event even over just subsequent weeks or months.
sorry, Mr. Anson, the man you are himself said no conclusion could be made about chronic exposure to parakquot and its association with Parkinson's. Anyway, let's um
>> but the authors in that study, I mean the hypothesis there was that MPTP induced Parkinson um may be manifested in only a few days post exposure when exposed at such a high level in a mechanism which is um replicable with some of the animal trials that have been done. Right. So
your hypothesis is that um after an acute highdose event within 30 days you should be able to see um some evidence of Parkinson's. Can you provide some evidence for that?
>> Um sure we'll we'll provide that. >> But I guess like my piece that I'm trying to avoid here is we don't look at each individual study on its own. It's awaited evidence of all of the studies that are produced. Every single study that was sent to us is considered interrogated further, but it adds to the weight of evidence and that averaging of all of that weight of evidence to arrive at a conclusion is what the key role in a regulatory sense is. So we can go through individual paper by paper and we intend in our submission to the inquiry to do that in more detail. >> But it's that average weighted weighted evidence >> that we use to arrive at a
that really stuck out reading um your decision is that you point out the shortcomings of studies that don't um back up your decision and then you don't do that with ones that do. Um and um one academic actually got in touch with me um and noted that studies forming the basis of the APMA's conclusion of no convincing association are largely funded by or written by people connected with Senta. Um Sme 2016 the author served as a paid consultant to Sententa. Um, Minimma 2014 funded by Sententa and authored by Sententa employees. Uh, Breenidge 2013 noted by the APV as being a study that was industry supported. Um, Vachary 2019 after the study was finalized, the author became an employee at a private consultancy for pesticide manufacturers. Um, and then I found this next one very interesting. Um, weed 2021. The author is from a private consultancy, not actually an academic institution. And his latest papers were titled, "Do cell Cell phones cause brain tumors? Another piece of the puzzle?" And a metaphor for epidemiology, the sea of person time? Um, his follow-up piece, Weed 2024, um, was published in an open access journal that's been flagged globally by academic institutions as predatory or questionable. Um, this journal charges $4,000 to publish an article, and you've relied on this evidence in your And apart from the Breenidge one, none of this was mentioned in your summary of studies. Uh why was there not a note as to who who funded these studies? That seems fairly standard practice.
>> And on saying that, sorry Senator, but uh Mr. Hansen, it is 9:15, so I'll let you answer that. Then we do have to take a scheduled break. So we give these poor hardworking devils behind us a bit of
fresh air. Look, so we're very conscious of the potential for bias in all sources of information that's used in the assessments and not just bias from the perspective of industry bias. And it's always a challenge for us given the fact that it's largely industry who are investing the dollars to do the more most detailed of the analysis and the studies and that's in any of the assessments that we do. But there's also inherent bias in the papers and in in the authors who are um disease litigation expert witnesses for plaintiffs in court cases. Like there's a whole series of inherent biases that we acknowledge and we look at when we're looking at the papers. But at the end of the day, it's the data and the data contained within those papers that we look at as well.
>> But you didn't acknowledge any of those.
>> Senator Senator, I did I did say we went over a little bit. It's not about us. about others that need a break, too. So, we will take it. You'll have the call when you come back, but I would ask that we've still got seven um entities to go. Senator Thorpe has some questions in this space, too. So, we can put our minds to that if we can get through them quick as possible. All right, let's have uh please let's have 15 minute suspension and we'll see you all back here at um 9:30ish. Thanks very much. Okay, in continuation,
continue what you were saying earlier. Do you accept that industry sponsored research tends to make conclusions that are favorable to industry and that that it should probably be noted when you're relying on those studies that this was funded by the manufacturer? [snorts]
>> So we are conscious of bias or the potential for bias and all sources of information that gets provided to us for assessments and we try to build consideration of that potential into our views. It it doesn't mean we discount evidence solely because it's industry funded or solely because it's um research has been done by um expert witnesses for court cases. We we still look at those. We take into account the potential for bias, but we still look at it and it's an an unfortunate scenario that we face in which it's often the industry studies that provide the greater access to the underlying data and the methodologies because it's obviously part of their business is actually collecting the information to support the registration or the continued application for those as opposed to what might be available in terms of funding and studies that get done by NOS's or others. Okay, I'll talk
a question notice about that. But um I wanted to ask about some evidence um from the Movement Disorder Society of Australia and New Zealand, which is essentially the the peak body for neurologists specializing in diseases like Parkinson's. And I thought I'd just quote what they have to say. A robust study of international epidemiological literature demonstrates a consistent statistically significant association between parakquot exposure and increased risk of Parkinson's disease. In the farming and movement evaluation FAME study, every Parkinson's case was individually examined and confirmed by movement disorder specialist neurologists. FAM established that parakquot exposure surgery with 2.5 times the odds of developing Parkinson's disease. High resolution spatial studies in California utilizing mandatory pesticide use reporting records demonstrate that living or working within 500 meters of paracquart application sites more than doubles Parkinson's disease odds confirming a clear dose response relationship with retrospective recall bias. How is it that our country's neurologists are telling us one thing and you're telling us another thing and that is such a divergence? You're telling us that there's no convincing association while our nation's doctors are saying the evidence couldn't be
>> Yeah. So that there's two distinct pieces of studies there that might just I'll try to be as quick as I can to you. But the fame study which is a really important and valuable contribution in terms of the research this was the work that was done by Tanner in 2011 >> and it looked at that broader that um US a health study that I I started to talk about before it looked at that and it looked for um for um detections of Parkinson's and um then did retrospective assessment as to tell us your history. Tell it what's your recall with regards to and it came up with a uh an odds ratio of 2 2.5. Um so really important piece of study. that same cohort, that same group of data had had work done on it just before by uh in a research paper by Camel Edel in 2007 and it did an equally important bit where it looked at um incident cases and prevalence cases. So in prevalent cases, it's where you do that piece about you look at someone with the disease and you ask them to tell tell you what they have been exposed to or what they've done and try to find correlation between that data and the disease they have versus incident cases in which you collect history and records about someone before they develop the That same cohort that Tanner used in the fame study was a cohort that was used by and and they found that study also found I think it was about a 1.8 um odds ratio. So an 80% increase chance in those who had already had the disease and were being asked to recall what their exposures were to things in those in which they had just recorded the information and then had subsequently developed Parkinson's. It was it was a one. It it was a no correlation piece and it brought in this importance bit about the the and it's not like it it's this recall bias that does come to play and that's why those two both looked at that. There was like four or five other really important studies that looked at that same cohort of people and asked things about whether they wore gloves, whether they didn't wear gloves, their genetics, >> looked at a whole series of individual pieces. That that that bigger cohort though ended up being that cheer that I I continue to struggle to pronounce. So he looked at 38,000 applicators and 28,000 of their spouses, followed them for 20 years, was collecting information and followed up with their health. like that is dealing with the same group cohort of people that all those other studies were but in a 10 time magnitude in terms of and and it was looking at collecting that information in cases in which the disease was not yet prevalent. It it it's a one that um identifies that uh a null conclusion. It it goes from a8 through to a 1 something about a 1.09 09 sort of odds ratio, >> but that they all add to that evidence that we look at. And it's that weight of evidence about
>> sure
>> 20 cases versus 490 cases. It's it's that kind of weighing up that the toxicologists, not the neurologists, but the toxicologists that work within APVMA and study that relationship between chemicals and populations
The problem is though this the problem is the inconsistencies for me because Paul 2024 didn't have recall bias and yet you discounted it. It was designed not to have recall bias and to establish an individual level exposure estimate based on distance between their home and their workplace and recorded >> paracord applications but the APVMA >> So you say well it had recall bias um can't use it. this didn't have recall bias can't use it. Um, but I just wanted to go to the
>> Sorry, can I just because I know I said I was going to kind because that was the second way you talked about proximity to places and this is this is not recall biases. This challenge you get with ecological fallacy about the fact that because something is close to something, it must have been exposed to something in a certain way. And and this is one of the problems when you don't have any measurements that actually allow you to work out what distance from something do you need to be for there to actually be exposure at what level and therefore what impact might that might have. I mean that was the I mean Paul like that study was another good piece of additional work. The problem is its title in which it calls itself ambient parakquot exposure actually didn't measure any ambient parakquotic exposure at any of the residents or workplaces or sources. They merely measured the geospacing from farms and um agricultural areas and associated the distance from that to potential pesticide use and proximity to therefore exposure. And and that's a really challenging piece to try to use in a regulatory sense to work out. Can you draw a a line between exposure and Yeah.
>> But that's that's not ecological fallacy. You're you're saying it has ecological fallacy, but it established individual level exposure estimates. And as as one scientist said to me, it does not contain ecological fallacy in the traditional sense. Ecological fallacy describes the attribution of group level associations to individuals within the group. That's not what that study did.
>> No, but it's that assumption that all locations within a given proximity from a point source of a chemical release will have the same exposure regardless of prevailing winds, topography, the point source, the nature of the chemical, the pattern of application, a range of other factors that determine all that. That is where and it might not be the same classical definition and so apologies I'm using the wrong [snorts] shorthand on that but that that applying a assumption that a whole series a whole group or a whole is equal to each of the individuals is what we're
>> So what do you want a double blind trial where everyone you know drinks like what what level of I want to take you to the statuto test. Um, section 5 A1B of the vet the AV vet code requires the delegate to be satisfied that use is, and I'll quote, is not or would not be likely to have an effect that is harmful
>> That that's right. And and if we find that we are satisfied, then we must continue the registration or must vary the labels. And if we're not, then we cannot. And that that's the test that we end up coming down to, right? So when we look at this and we ask ourselves the question, do we think it's likely on the weighted evidence we have at the moment? And and I I say at the moment, the weight of evidence we have at the moment, the answer is no.
>> So do you do you need a double blind trial? Like what are you what are you waiting for? 70 countries have done
this. >> Um and the 70 countries was like I had to look up some of those countries you dropped on me last time to find where they were on a map. It it like those 70 countries are a really it's it's a a misleading piece that gives people false hope that there is a single silver bullet on this stuff. I mean th those countries you've got 27 EU member states that have not continued with paragot and and that's because of the EU assessment of deficiencies in the regulators process not not actually finding any health but concerns but their process was found to be deficient they sent it back to be reworked and no one decided to rework it. We we've got a whole lot of um um United Kingdom, Norway, all all around high acute toxicity like like this is a really hazardous chemical like you know exposure to it dermally. So exposure to skin is nearly fatal. Drinking a teaspoon of it is irreversibly fatal. Many of the countries in that 70 have removed it from their countries because they don't have the same mechanisms around poison scheduling control of use that we have here. And as a result, it it it was a both accidental and deliberate poisoning risk in those local conditions. And that's why you've got a long list of countries in the Sahil regional ban, the Gulf Cooperative Council ban. Like when we actually look at regulators who have talked about um potential linkages between Parkinson's and Parakquat, the the the one who talks about it but doesn't use it as the trigger for the the their ban was Brazil. out of all of the 70 something that are talked about the the other regulators in the OECD world who have done this the regulatory science risk assessment on this whether it's the Americans, New Zealanders, I I I I can't answer your question about what do we need to be satisfied because one of those things it'll be the weight of what we see in front of us. It won't be one study. It it'll be the next time we look at this and on the back of additional information that is generated.
have to share the last one? >> Yeah, thank you.
Paul 2024 has been described to me by scientists in this space as a very high powered study. So, I'm just like what kind of study would you actually need and and what what ethics committee would ever approve
>> Yeah. Again, I I I don't know how
>> we just
>> I don't know what it would look like. Sorry, Senator.
>> All right, Senator. >> Thank you. >> Thank you, Senator Thorpe.