This content is current only at the time of printing. This document was printed on 28 March 2017. A current copy is located at http://apvma.gov.au/node/19431
You are here
Transcript for Dr Jason Lutze, APVMA
Veterinary drug residues and changes in dietary risk assessments
This presentation was delivered at the APVMA’s science feature session on 15 October 2015. The full video is available on our YouTube channel.
There is a little bit of double‑up or triple‑up in our talks today and I think that's really important and each of the presenters is covering this from a slightly different angle so I think that is really good. I am going to talk about veterinary drug residues and development in dietary risk assessment for vet drug residues. I'm going to be taking a view here, especially considering the changes that we've had in the APVMA in the last 12 months, taking a view of where we have been, where we have got to and where we might be going in the future, particularly in light of some of the things that are happening in the international community at the moment.
First off, what is a dietary exposure assessment? We have already heard from Biraj and from Janis that it's an assessment of the risk that is associated with residues in food. It of course considers both hazard and exposure, and as the comparison of a health standard or a health‑based guidance value with the potential exposure. This exposure is related to the expected residue concentration in the food and the consumption of that food. There are two basic assessments we undertake depending upon the nature of the toxicity of the active ingredient that we're talking about and that is, the chronic or long‑term exposure assessment and the health standard or health‑based guidance value for that is the ADI or acceptable daily intake. The second is the acute or short‑term exposure assessment and the health standard for that is the acute reference dose. We've heard a lot about those and occasionally, very occasionally, we can come across some compounds where the toxic response isn't strictly chronic and isn't strictly acute and in those cases we've got to be really careful with how we parameterise the models and what exposure data that we use.
Most of you are aware that there are some differences in approaches, internationally differences in approaches, between pesticides and veterinary medicines. Before we get into the more scientific part, Phil talked this morning and Phil has just run out for his parking meter, he hasn't run out because he doesn't want to hear what I'm going to say. We've got to think about the science that we do in the regulatory context. What are we actually doing in an exposure assessment? We've got to do an exposure assessment to be satisfied against our safety criteria. We need to ensure that the use that we're approving would not be likely to have an effect that is harmful to human beings.
The legislation provides us with further guidance on this in Subsection 1A of the Schedule of the Act. It gives us some guidance on how we need to go about these assessments. In doing this, if we would immediately just think about safety criteria and that it would not be likely to have an effect to human beings, well, we could take an absolute zero risk approach and not approve anything but we're not allowed to do that. We need to recognise that agvet chemicals are important for our primary industry but as part of that we are also told that the driving factor that we need to consider is the health and safety of human beings. That is really at the top of the pile in the things we need to consider, that our approach and our assessment is conservative and that human health is being protected.
In doing that, the practices that we need, need to reflect internationally best practice and be based on risk in science. They also need to balance regulatory effort and promote community confidence in the regulation of agvet chemicals, and that's not only important to us in defending ourselves, it's important to those selling treated food and importantly our agvet manufacturing industry. We need to be conservative in our assessments but we only need to be appropriately so.
Again, we've talked about the chronic assessment, which is effectively ensuring that the amount of chemical that we consume on a daily basis for our entire lifetime is acceptable. An acute assessment, that's looking at the hit of residue that we might be able to take in a single day without our health being affected.
A brief history of exposure estimation with regarding to vet drugs at the APVMA. When we were established as the national registration authority, for vet drugs we used the JMPR methodology and I'll talk a little bit more about that in a moment. It's quite similar to what Janis was talking about earlier for chronic exposure. A review of the way veterinary residues were approached at the APVMA was undertaken in 2004. Following, a recommendation came out of that, that we adopt the JECFA methodology. Now through this process or just before that review was taken into account, the acute reference dose and acute methodology was developed and Biraj has talked about that, and that has been applied to veterinary medicines in Australia as well through that time.
About 18 months ago, coming up to two years ago now, there was another review undertaken by the APVMA on our implementation of the JECFA methodlogy and I'll talk very briefly about that paper as well, and some developments at JECFA that are looking at ways that the changes in chronic approach might take place.
The JMPR methodology
I guess the first part of this is the establishment of a residue definition. We've heard about residue definitions and we often establish two. The residue definition must be simple enough to establish a standard, establish an MRL to enable us to effectively monitor the use of an agricultural and veterinary chemical. Early on under the JMPR methodology, I guess the residue definition that is normally established is somewhat analogous to the marker residue under the JECFA scheme.
If that residue definition that we're establishing for compliance doesn't account for enough of the toxic residue that's left behind, enough of the toxic metabolites of the compound, we may need to establish an additional residue definition that is used for dietary exposure assessment. Under the JMPR methodology the residues of the preferred withholding period are then subject to a dietary exposure analysis based on the Australian food consumption data that Janis was talking about. If acceptable, this is used as a basis for establishing MRLs.
Just briefly, we've got a residue decline curve, the median we've got a distribution of residues at each of our time points and in this plot the food withholding period is here at time two. We look at the median residue and if that's acceptable from a chronic viewpoint for exposure to the Australian population, we can then use that to establish an MRL for an appropriate and inappropriate level.
The JECFA methodology, as we've discussed in the review in 2004 into the Residue Processes of the APVMA recommended the adoption of the JECFA methodology for vet meds. Since that time, this methodology has been used to establish MRLs. This methodology uses a model diet to estimate MRLs on the basis of acceptable total residues so this is one of the key differences between these two techniques. This technique assumes that all the residue that can be identified in the trials is actually of toxicological concern. Now, the review recommended that the exposure assessment should actually continue as before on the basis of Australian food consumption data. Now, what were the primary reasons for the independent review to recommend that we should adopt this methodology? Firstly, it was considered that we might end up with MRLs that are more harmonised with those of our trading partners and particularly CODEX, the EU and the USA. Secondly, there was a thought that us adopting this methodology might give us more influence in the CODEX is the CCR BDF.
How does this work? I thank Phil Reeves for this short series of slides. This is very complex and I'm not going to go into this in a great deal of detail but firstly, the metabolism data is used to identify what an appropriate marker residue should be. That's the residue that at the end of the day we're going to establish our MRLs on. Then we use some residue depletion studies using labelled or hot residues to establish how that marker residue relates to the total residue in time. At each of our three time points we've got a residue profile for marker and total.
Then we go through. Now, when applying the true JECFA methodology, we're thinking about the first time point that we should be thinking about is something that equates to the good veterinary practice associated with the use of this drug. This is something that differs greatly between different regulators, especially with regulators where there is a large separation from risk assessment from risk management. There might actually be no consideration of the actual use that is associated with the chemical when the MRL was actually being established.
This is an iterative process. We'll go through and look at each of these time points and let's, for instance, say that at this level in this instance if time one was the preferred withdrawal period or withholding period for this veterinary drug, we would then look at the total residue occurring at that time one and decide whether on the dietary risk assessment whether that was acceptable. If it wasn't we would then fall down to the next time that we have data for and we would undertake an iterative assessment at that second time point, and if that's acceptable we take through, go through and establish an MRL at that point and proceed to approve the application.
However if we get to the endpoint and it's not acceptable on the basis of total residues at any of these time points then we'd have to draw a halt and say, "No, we can't support that use." I'm not going to defend this methodology at any way at this time. I'm here to talk about the assessment methodology for dietary risk rather than the MRL establishment at this time. There is a large amount of conservatism built into these processes. It's there because the dietary exposure methodologies need to be conservative to protect public health but they need to be science‑based. In these dietary exposure assessments where is the conservatism? Well, it really lies in two places and that is, the residue assessments that are fed into these calculations and also on the food consumption estimates.
Firstly, if we look at residue estimates. For chronic estimates we assume that all food consumed has been treated and contains residues at the median or in some cases high level observed in the residue trials. Now, particularly under the JECFA methodology, we're not just talking about the residue that we know to be toxic but we're talking about the total residue that's been identified in the trials. This is really conservative and this is some Australian data out of the National Residue Survey, the random programs out of the Residue Survey. There's some data here from a ten‑year review of some anti‑benthics, which are quite commonly used in Australian practice, and we can see for the benzimidazole, for compounds such as thiabendazole, in over 1700 samples of cattle analysed over those 10 years there were only three positives. Resides of that group of compounds were only identified in animals at slaughter three times over that 10 years.
For the macrocyclic lactones, there were about 3,000 samples collected over that period and there were only 55 positives. That hasn't changed a lot. This is some data from the last 12 months. Very similar for the benzimidazoles. There's a little bit of an increase in the macrocyclic lactones because at about the time that that analysis was done there was a change in tissue type that was being analysed which increased the sensitivity of the survey. Assuming that residues are going to be present at the median which is the point that we use for chronic assessment here in vet residues is always going to be present at the median is really very conservative. Other regulators even consider that that median isn't actually conservative enough and they will use the MRL or higher residue observed to do their chronic assessment.
Then we move over to the food consumption. The check for process for estimating MRLs uses a model diet rather than real world food consumption data for the market of interest. This assumes that essentially every day of our life we are going to be eating 300 grams of muscles, 100 grams of liver and also 50 grams of kidney, 50 grams of fat, 100 grams of eggs and a litre‑and‑a‑half of milk, and that we're going to be doing that every day of our life. I think sometimes we might get to 300 grams of muscle in a day but I know once I've had one of those big steaks I don't really feel like eating much for a couple of days after that. How does this really compare with the data that FSANZ provide to us from an Australian viewpoint?
On a population average basis we actually eat about 160 grams of muscle a day. For liver and kidney we don't eat very much at all so for our toxicity profile that's truly chronic, consumption of offals in Australia for the general population isn't of high importance. We eat about 10 grams of fat, about 15 grams of egg and actually about 740 or 750 mls of milk. That's dairy product coming from all sources amplified up to milk as Janis discussed. So the JECFA model diet is really very, very conservative and we are continuing, although we're using the JECFA process and the JECFA food consumption data for estimating MRLs where it's appropriate, for our food actual exposure estimates we're continuing to use Australian consumption data.
We have talked about this a little bit. Chronic exposure assessment in Australia for vet drugs is very similar to the FAO/WHO methodology where we look at the sum of the mean consumption and the median total residue for all foods to compare that with the ADI.
I'll move on to short‑term exposure assessment. We've talked quite a bit about this in the last hour. Acute toxicity has been considered internationally and in Australia since about 1997, and more recently by JECFA. In Australia and increasingly so internationally, for compounds that are considered to be acutely toxic an acute reference dose is established. Biraj has gone into this in quite a bit of detail but last review that's been finalised into these equations was undertaken in about 2005 and the IESTI was put in place. For animal commodities this is taken as the large portion, the 97.5th percentile large portion against the high residue or the high total residue in veterinary medicine's terms or the median residue where appropriate for items such as milk, but even some regulators they don't consider a median residue is appropriate for milk and they will use the height of the MRL. They adjust it for body weight.
In the short‑term exposure assessment this differs somewhat from the chronic in that we're only considering one food at a time, whereas in the chronic we're considering that we're potentially exposed to more food sources. Children and the general population are considered separately and this procedure has been applied by the APVMA to veterinary medicines since its adoption by the APVMA in about 2000.
Developments at JECFA
JECFA has not often explicitly considered acute toxicity. I guess there are a couple of reasons for this. They consider that the TMDI methodology that they use has been very conservative. The JECFA workshop in 2011 considered new methodologies for both acute and chronic dietary exposure estimation.
Since I have got two minutes I'm going to skip past the acute one because that hasn't changed very much except it's looking at using some local consumption data in the same way that we do here. They also looked at changing chronic assessment methodology because they wanted to move away from the model diet to more realistic diets and to incorporate other tissues for where necessary where some populations might consider foods that aren't really considered in that model diet. But then the meeting went on to consider that mean food consumption was not conservative enough to account for habitual high consumers. They developed an equation, which is called the global estimated chronic dietary exposure to encompass the risk to high‑level consumers of foods. This includes an estimate, which is akin almost to the acute calculation where it uses 97.5th percentile consumption for a single and the median total residue for a single animal‑based food, which is going to produce the highest estimate of this residue. Then it sums that with the chronic calculation for the remaining estimates of that animal‑based diet. If liver is going to give us our highest estimate from this part of the equation, we would calculate both the high level exposure for liver and sum it to the mean and median consumption for milks, muscle and fat.
Where does this leave us? This is currently being assessed by JECFA and CCRVDF. In a number of instances out of the last meeting, this estimate actually resulted in more conservative or higher exposure estimates than the model diet. Something that the committee is currently looking at and we are looking at as well is, is this further conservatism really necessary? When we look at the things we talked about before with the level of residues we actually see in the population, at the moment we don't actually think this extra conservatism is really necessary at a national level.
I just wanted to touch on this review of JECFA methodology and there are two really key recommendations that we are implementing at the APVMA at the moment. This review was done including submissions from Animal Medicines Australia and was undertaken by three international experts. The two key recommendations will be that we should continue to apply the JECFA approach but to apply it in a way that it was designed, as a tool and not a rigid rule. That we use the weight of evidence and total and consideration of all the data that is available to us. I guess secondly, that we actually apply appropriate regulatory process to vet residue assessment and dietary exposure assessment so that we're looking at acceptable risk rather than nil risk. If any of you would like to talk to me further about that and what we do in that regard, I'm quite happy to talk to you about that.
Where are we going?
We are going to continue using Australian consumption data for chronic and acute exposure assessments. We're going to watch the further development of the new JECFA methodology. At the current time we don't consider that implementation of the modified chronic exposure methodology is needed but we are going to be running some trials of that new methodology and the evaluations we've got, and if any of you have got other residue datasets that we are not likely to see in the next little while formally, I'll be quite happy to run that new model through that dataset as well so we can be more informed when we go into CCRVDF next year.
Thank you very much.
Errors and omissions excepted; check against delivery.