Animal & Veterinary
Session IIIb Questions and Answers
DR. ROACH: Yes, Steve Roach from Food and Animal Concern Trust. Doctor Silley, I thought your description of the problems of the surveillance system were very good. And I’ve tried to work with some of that data before and I have the same problem in making comparisons. I think the problem where you mix clinical and slaughter isolates together like in DANMAP, causes problems, because it just makes it very difficult to know what’s happening.
But I just had a question or a comment related to your statement about the high level of cephalosporin resistance in the poultry industry. You know, for years I had the same problem that you had, it seemed really odd that you had this -- where you’re finding a lot of the CTXM, or the extended spectrum data --- in the poultry it wasn’t approved for there, but I think at least from discussions I’ve had with some of the people looking at risk management in Europe, is that in fact there was quite a bit of use and probably is still ongoing on use of cephalosporins in broiler hatcheries. And as somebody from IFAH you may have some information on that as well.
DR. SILLEY: I guess I can only really comment on the fact that there aren’t any approvals. If it’s used then it’s clearly it’s off-label use but I’m not a veterinarian, I’m a microbiologist, and I’m simply actually looking up the data. Clearly it must be coming from somewhere. You know, that’s a reasonable assumption, but the actual data of where it’s coming from remains to be seen.
DR. ROACH: I think, at least the information I’ve read, at least the broiler industry in the Netherlands hasn’t --- having you said it, have agreed to ceased to use it, so -- so, there is some information besides just speculation on where it’s coming from.
DR. M’IKANATHA: I just want clarification, when you showed isolates, clinical isolates and wild isolates, were you comparing across the species or were the clinical isolates from pigs versus the non-clinical isolates from pigs?
DR. SILLEY: Yes.
DR. M’IKANATHA: Which one is which? Are you talking about human isolates?
DR. SILLEY: No, no. All the data that I’ve actually shown you today is all been animal data.
DR. M’IKANATHA: Oh. But the temporal relationship, at least here in the U.S. that we conduct surveillance on, comparing the element of resistance within the isolates that in for example poultry, then the temporal is only human isolates not on the poultry isolates. But anyway that’s number one.
Number two, the problem you showed about the change of cut-off point that is very well known that in surveillance if you change the case of definitions you must apply the data that were being analyzed using the old case definitions and then cutting forward showing that the other data you analyzed you are using now a new case definition.
I don’t think there should be, and you might consider that, any need at all to say you have to standardize data across the countries. You can if you start to use a new surveillance system, but each country I think the data will vary and what is important at least in surveillance is looking at the trend over time. So even if you’re using a different cut-off point and I’m using a different cut-off point, as long as we’re looking at the trend, the data should still provide useful information. Thank you.
DR. SILLEY: That’s absolutely correct. However, if you actually don’t understand that and if you simply go to a report from a national surveillance program, and you look at the headline resistance right, and you don’t go any further, then actually you’re not comparing like with like. And I agree with you completely. But I think if we’re actually looking to harmonize so we talk a common language, and I applaud what EASSA is trying to do, in the presentation that we had earlier.
The only concern I have there is whether we’re actually using the right cut-off values because for example if you look at the EASSA report and you look at the data for the Netherlands, for Salmonella resistance, E. coli resistance, and you compare that with the MARAN report for the national surveillance, you get two different figures. So, it is somewhat confusing.
DR. LYNFIELD: Ruth Lynfield, Minnesota Department of Health and IDSA. I was not aware actually about this epidemiological definition for non-susceptibility. And I’m wondering if you can answer a couple of questions for me.
Is the -- you show some nice graphs where there are clearly was a difference in populations. It that always the case? Is there a general approach that you take by using a proportion, something doesn’t make the 90th percentile, then it goes over in the other camp? And what are the data that support an approach that you have a wild-type if it is a certain proportion versus another?
DR. SILLEY: A lot of it depends actually on the resistance mechanisms that for the particular drug. Because if you take the easy one with fluoroquinolones and you got first and second staph mutations, then you can clearly define those things. That’s pretty straight forward.
But I think there’s a lot of work that still needs to be done. And in fact, the presentation that Jeff Watts gave from CLSI and the XO8 report, is actually trying to grapple with that. And actually -- to actually link how we do define those wild-type cut-off values.
DR. LYNFIELD: So, I guess I wonder why do we need to do that. Why not use the clinical breakpoints. If we use it for human medicine and this way we can compare what we see in human medicine, we’ll be using the same language.
DR. SILLEY: A very quick answer, we’re running out of time. I agree with you completely, why can’t we use the clinical breakpoints.
DR. CARATTOLI: Can I comment? We don’t use it because it --- micro organisms that adopted the bacteria starts with a very low MIC, that this means that they are prone to develop, to survive this special environment like the intestinal track, and so they can develop resistance very soon. So with the speed and with the positive selection that you don’t expect. So what can we do with this.
DR. SILLEY: And I think that’s right. And I was sort of being a little tongue in cheek when I said yes, let’s just use clinical breakpoints. I think the issue is depending what is the question because there are two separate things. The clinician needs to know about resistance from the clinical perspective and therefore is the drug that he’s actually prescribe got a chance at working. The epidemiologist and those of us interested in actual resistance development actually need to actually look to use that fine turning to actually see those things.
So, I made the point that the epidemiological cut-off value is a value, it clearly is. And I think what we should be looking at is both. We should be using both.
My big concern is don’t call resistance that which is determined by the epidemiological cut-off value. It’s decreased susceptibility, it’s not resistance because the public at large relate resistance to something that’s actually not going to work. And I think the great danger is we’re not communicating the right things to the people outside.
So, I would like to use both and clearly determine decrease susceptibility as being different to clinical resistance.
DR. LYNFIELD: Well, the one last comment I want to make, I mean obviously it’s a very complicated area, but that’s what I am concerned about, is there are so many different resistance mechanisms. There are also different ways of conducting susceptibility testing. And if you’re just looking at the MIC, is that MIC from a broth micro dilution, is that from a disk, and that -- or is it from an E-test, they all may have some different breakpoints as well. So, it is a tricky area.
DR. McDERMOTT: Just in case you thought we might be getting closer to harmonization, I wanted to point out that in NARMS when we don’t have clinical breakpoints we set our cut-off values at the lower end of the resistant population, just so that we can have even less harmonization perhaps. But really it’s because FDA incorporates CLSI standards by reference so we’ve always had a preference for the CLSI approach to defining resistant populations.
So, if you look at say our Campylobacter where there are no clinical breakpoints that have been established in the standard way with the three datasets, we use --- breakpoint at 32 which is right at the lower end of the resistant breakpoint.
It wouldn’t matter too much if we use the top end of the susceptible breakpoint but in some cases, as you point out, it could make a big difference.
I vote for non-wild-type if we want to vote on terminology because I think you imply even with decrease susceptibility only a lab -- perhaps only a laboratory measure. But harmonization is a huge issue, Peter, as you pointed out.
DR. SILLEY: And I just think one final comment, I think it just invites the point that Pat made in answer to a question earlier, and the only reason that I was able to actually do some of this analysis was because the national surveillance programs in Europe actually publish, largely publish susceptibility profiles. So you can actually see the susceptibility distributions. And if you got the distributions you can actually start doing the analysis however where you want.
My big concern really is this use of this word resistance and we’re very careful about how we use, or we should be careful about how we use it. And I believe very strongly it should be simply related to clinical resistance and we should use something different in terms of that which is branded by the wild-type distribution.
But I would encourage people to publish that sort of susceptibility distribution and I think by doing that then we’ve actually got a way forward. And which is what Pat suggested earlier. And thankfully most of our European surveillance at least does it that way, and that’s a very positive thing.
DR. CARATTOLI: Thank you very much, Peter. So, this session is completed now. We can change. Thank you.