GRANT L. CAMPBELL: This panel session is entitled Laboratory Testing for West Nile Virus. We have five excellent panelists with us today. My co-moderator, Dr. Leo Grady is from the New York State Department of Health. My name is Roy [Grant] Campbell from CDC in Fort Collins, Colorado.

LEO J. GRADY: Good morning. Sitting directly to my right is Dr. Cinnia Huang from the New York State Department of Health; to her right is Dr. Laura Kramer, also from the New York State Department of Health. Dr. Rob Lanciotti from the CDC, Fort Collins, is next, and then comes Dr. John Roehrig, also from the CDC. At the end is Dr. Bob Shope from the University of Texas Medical Branch in Galveston, Texas. Each of the panelists will make a few brief comments about their involvement in laboratory testing, and then we will open the floor for discussion. Dr. Huang begins this session.

CINNIA HUANG: Our lab is involved in human testing. The new state experiments with RT-PCR for West Nile virus are similar to what has been done at Fort Collins. The only difference we find is that a two-step regular RT-PCR is as sensitive as Taq-Man. The two-step RT-PCR is the protein protocol used now for the encephalitis PCR panel. Our panel includes different viruses, and West Nile is one of them. Based on our data, there are cases other than West Nile virus, such as enterovirus and the herpes virus, and the enterovirus season overlaps with West Nile virus. For this reason, our lab will continue to perform tests using PCR batteries.

ROBERT LANCIOTTI: I want to run through the most commonly used diagnostic tests, both for serology and for virus detection. For human diagnostics, our gold standards are the IgM and IgG ELISAs, and we confirm those by plaque-reduction neutralization testing. The molecular amplification tests, such as RT-PCR and TaqMan have limited usefulness in human diagnostics, and I think it’s important to emphasize that. There are some people with a misconception that those could be front-line tests for human diagnostics. We have data showing—and it’s limited data from the 1999 epidemic—that a positive PCR certainly is diagnostic but a negative doesn’t mean a whole lot. There are a lot of other reasons. We have serologically confirmed cases, where you can’t detect viral RNA, and I think that has a lot to do with the duration and level of viremia. Many times we’re looking for virus pretty far out beyond the febrile phase, so the molecular tests have been most useful in the fatal cases, less useful in CSF, and even less in serum. Again, the ELISA serological tests are really our gold standard for human diagnostic testing. Virus isolation also should be attempted, but in many cases, the virus samples that we’re getting to test and isolate are, in the diagnostic lab, beyond the viremic stage in many cases. But I think they’re still important tests and ought to be attempted. For the surveillance of mosquito pools and birds, this is an application where the molecular tests are very useful—the RT-PCR and TaqMan. NASBA is another amplification technology that works very well. Those are the most commonly used tests, not only in our lab, but also in other state health departments. There’s also immunofluorescence and immunohistochemistry, and there was a poster here describing antigen capture ELISA for West Nile detection as well. I think Dr. Kramer is going to talk about the importance of having a confirmatory algorithm in the lab. I think for any one of these tests, whether serological or virological, there needs to be a way of confirming positives. The other issue that I’ll just mention is the importance of standardization. You may be aware that the CDC has done a lot of training in the last few months. We’ve trained, at this point, representatives from almost every state health department in the country, and we are now initiating a huge proficiency panel program where we’re sending out both molecular and serology samples to every state health department that chooses to participate. This will be a large undertaking, but we think it’s very important for states that anticipate testing to participate in the standardization proficiency panel.

LAURA KRAMER: To follow up on Dr. Lanciotti’s comment, I think proficiency testing is crucial, especially with all the new labs coming online now doing surveillance testing. I also think that there’s got to be quality control to make sure that what is positive is really positive. Because PCR is such a powerful and sensitive tool, it’s a double-edged sword, and it’s very easy to pick up contaminating RNA, either by aerosol or on your glove when you open up a tube. It’s just so sensitive. We know that it detects less than one plaque-forming unit. We diluted avian tissues out 10−6, and we still detect a positive. So, it’s very easy to think that you have a positive tissue and not really have it because your RNA is contaminated. I think some water blanks need to be included in all the assays, particularly with the bird tissues, where there’s such a high load of viral RNA. The water blank needs to be in there from the very beginning. Where you’re homogenizing the tissue, an aerosol is being created; it’s closed, the system is closed, but when that tube gets opened it’s very easy to carry over some viral material and think that you have a host of new positives. So if your water blank comes up positive, whether it’s from the beginning or whether it’s from the RNA extraction stage, you need to go back to the tissue and reextract and repeat the process to confirm that that tissue is truly positive. And as part of this, when labs are setting up new TaqMan or PCR assay rooms, they have to be fully knowledgeable about how the flow needs to go. There are many books on it, but if you’re doing standard PCR and writing gels, you never take those pipe headers from that room after you clean your area, which is your pre-PCR area. For a confirmatory assay, just to reiterate what Dr. Lanciotti said, there needs to be an independent confirmatory assay.One set of primers may be more sensitive than the other, but we don’t know what you’re detecting if you keep on repeating that set of primers and saying, “Well we did two assays,” when, in fact, it’s really one assay repeated. It needs to be two totally independent assays. So, either two sets of primers or an RT-PCR and an IFA or a NASBA—some other totally separate assay. I think that while RT-PCR is wonderful and it’s rapid and you can get the results out there, I don’t think we should stop trying to isolate virus and cell culture. First of all, you can pick up viruses that the PCR is not specific for. PCR is a very specific assay because of the primers, and with the TaqMan probe, it’s even more specific. But, it’s important for research to understand the virus and for confirmation that this is an infectious virus to inoculate cell cultures. What does a PCR positive mean? I think we need to try to sort out what’s happening with the mosquito pools and the bird tissues as well, where we do see RTPCR-positive tissues that we cannot get out in cell culture. And I think we need to determine how we’re going to interpret these.

JOHN T. ROEHRIG: I have a few comments. It’s come to our attention that some of you are setting up BSL-3 laboratories in which you want to do plaque assays for serology and also virus isolation. That can be a dangerous approach to take if you’re trying to do isolation in the same place that’s using standardized West Nile for plaque assays because you’re liable to get cross-contamination. Last year when we talked in Fort Collins about the best approaches for RT-PCR, one of the suggestions that we had was that for the plaque assay it may be useful to use a prototype strain like Egypt 101. In our hands it looks as though the New York 1999 strain and the Egypt 101 strain are pretty interchangeable. If you were to use a strain like Egypt 101 in your plaque-reduction neutralization test and happen to get that cross-contaminated into a tissue that you were trying to do isolation from, you could actually tell those viruses apart by sequencing the amplimers by RT-PCR. Unfortunately, at the time, everybody demanded New York 1999, but now that we’re a year and a half out, you may want to reconsider how you’re doing that plaque assay and go over to something that’s actually distinguishable from what you might get from isolates from the environment. My history in arbovirology has been less on nucleic acid and more on proteins and protein structure and function. I just want to remind you that there’s another component to viruses besides nucleic acid and it’s actually protein that can be used as a confirmatory test. Laura’s laboratory uses IFAs on tissues and isolates. There are plenty of monoclonal antibodies available to you that will allow you to identify West Nile and other arboviruses that you may be isolating or seeing. Yesterday, there was a poster on an antigen detection assay that’s been developed for West Nile that’s as sensitive as the SLE assay. It’s been out for a number of years and can be used in tandem either with mosquito pools, or, in this case, it was also shown in tissues of birds, where luckily the virus titers are even higher than in mosquitoes and a relatively good assay. It could be used either as a front-line assay or as a confirmatory assay for other tests. So, be aware that that’s out there and should be available to you if that’s what you desire to use as well. Later on today or tomorrow, Denise Martin will talk about some of the work she’s done looking at the cross-reactivity of IgMs between SLE and West Nile. I’d like to remind you that there are other things out there besides PCR, although PCR is quite good. But, there are other things out there besides ELISAs that you can use to identify antibody reactivities in other species. But you have to be careful when you do things like hemagglutination inhibition assays on birds. The extraction protocols for the sera are a little different than what you may be used to as far as the other nonspecific inhibitors that may be in bird sera that are not removed by the standard sucrose acetone extraction. Bear in mind that there’s a lot of literature out there and, you know, one of the implications that’s come out of this meeting is that there’s a lot of new information that’s coming out that people haven’t thought about. I think lots of people have thought about these issues for a long time before we’ve even thought about them. Nothing substitutes for going back and looking at what’s been done in the past and seeing how you can apply it to the present and the future.

ROBERT E. SHOPE: Let me first agree completely about the use of Egypt 101 as a strain for neutralization tests for the reason that you’re giving but also because I think you’ll find that it produces much nicer plaques than the New York strain does. I don’t know why, but that’s a practical reason to use it. I’d like to make some comments about cross-reactions among flaviviruses. I first learned about the cross-reactions from Jordy Casals, who incidentally just lives down the road here in northern Manhattan, but he described the cross-reactions and flaviviruses using the hemagglutination inhibition test and later using the neutralization test. What happens is that in a primary infection the serological reaction is relatively specific during the first week or ten days postinfection, but then it broadens out and you can use any one of the serological tests against any one of the flaviviruses to show that the person has had a flavivirus infection. As in 1999, it was discovered, even with the capture IgM test, which is relatively specific, that you can misdiagnose closely related flaviviruses. What was called St. Louis encephalitis was really West Nile, and this isn’t surprising. This is what one would have expected. I don’t have the benefit of working with very many West Nile human sera, and I look forward to hearing the talk tomorrow, but in our animal models—we have an animal model of West Nile—Bob Tesh has developed a model in a hamster that, as far as the clinical syndromes go, is very close to the human picture. About 40% of these animals die following an interperitoneal inoculation. These animals develop a muscle weakness prior to death, and they also develop serological reactions by day six after inoculation; they have rela tively specific antibodies in the hemagglutination inhibition test, and they also start to get IgM. The IgM is gone by day 60, which I think is probably very close to the picture in humans. Is that right?

ROEHRIG: I should have mentioned that. Actually, it goes a little bit longer than that.

SHOPE: OK. But it certainly decreases rapidly just before that time. Regarding the secondary infections, if you inoculate a hamster with yellow fever or Japanese encephalitis virus, follow it and a month later or so inoculate with West Nile virus you get a very brisk anamnestic reaction, which is broadly cross-reacting in all of the tests that we use, so that you need to be careful how you interpret results. Bob Tesh and and his colleagues have been testing human serum in a study collaborating with the Department of Defense using recruit sera from the postal codes zones of the epicenter of the 1999 epidemic. There are no identifiers on these sera. Other than the fact that they’re probably eighteen, nineteen, or twenty year olds, we don’t know much about them. But we do know where they were recruited from. They’re using an HI test, a hemagglutination inhibition test, and many of these sera have positives. Many of the reaction patterns when you test with multiple flavivirus antigens look either like dengue or like multiple reaction sera. These people were bled before they were vaccinated with yellow fever, but some of them may have had prior vaccinations also. So the message here is to be very careful about interpreting serological reactions. A positive with West Nile does not necessarily mean that that person has been infected with West Nile virus. This stimulated us to look at some of our dengue sera. We have four prototype sera that were supplied through Jim LaDuke from dengue patients. There’s one from Malaysia, one from the Philippines, one from India, and I forget where the fourth one was from, but we went back and did both West Nile IgM ELISAs on those sera and they’re positive, in addition to being positive for dengue. And Bob Tesh also went back and did neutralization tests. Three out of the four sera are positive in the West Nile neutralization test. He uses a 90% plaque reduction, so it’s a very stringent test. So, the message is to be careful. In the hamster experiments, we’ve done a vaccination challenge. The vaccine that Bob Tesh used was the live attenuated Japanese encephalitis vaccine, the vaccine that has been given to millions of horses in China. Those hamsters did not get sick from the vaccination. They were challenged a month later with West Nile virus, and in the controls there was about 40% mortality. There was no mortality in the vaccination challenge test with West Nile virus. So here’s an example of a vaccination challenge test that also shows more cross-reaction. There are other implications in the vaccination challenge experiment. I mean it must be obvious to you that the vaccine for horses used in China probably would be a good vaccine to use for horses in the United States.

MIKE L. BUNNING (Centers for Disease Control and Prevention, Fort Collins, CO): Can I ask John to follow up on the IgM persistence?

ROEHRIG: Sorry, I didn’t get that written down on my list. We’ve been collaborating with the New York City Health Department doing follow-up studies on the longevity of the IgM from some of the patients from the 1999 outbreak, and it’s been pretty revealing, although not totally surprising about the longevity of the IgM in some of these patients. Clearly what Dr. Shope said is true: if you look at the IgM response early in infection, and some of these were pretty early infections in 1999, there’s a clear spike of IgM activity. This is serum now we’re talking about, not CSF. In the early serum infection after 30 to 60 days, it will drop down to lower levels. However, many of these patients carried IgM and continued to carry IgM in their serum specimens above the cutoff level for the IgM ELISA. Not many at this point. I think the last lead was around 500 days postinfection, and at this point the study has been terminated. But it was a very valuable study. What we were interested in looking at and why this was so important from a public health perspective was whether or not IgM could persist from one transmission season to the next at a level that if you happened to have an individual from outside the United States come in with encephalitis the following season, and you did an IgM on them, or something was suspicious of encephalitis, and the IgM was positive but at a lower level, is this truly a recent infection of West Nile or a past infection of West Nile? I think our conclusions are that you actually can be confused by that. That’s why earlier on this season, we were recommending acute and convalescent specimens that could be tested as paired sera to actually sort out whether it’s an active immune response or if it was a stable older immune response. So it’s clear that some patients can carry IgM from one transmission season to the next. Up to this point I think that six months was about as long as this had been investigated, although never with West Nile. With Japanese encephalitis it had been looked at, and actually I guess with West Nile with some of the Romanian patients, Dr. Tsai looked at six-month specimens. So bear that in mind. As people who do serology will tell you, it’s clearly always better to have acute and convalescent specimens that you can actually titer. This especially becomes important when West Nile travels into regions where there is more activity from St. Louis encephalitis and you may actually be looking at an SLE infection versus a West Nile infection.

CAMPBELL: So we know now that, at least in a few patients, that IgM would persist for up to a year, and that might have a couple of implications. It goes back to something that Dr. Shope and others have said and that’s to be careful because you’d have to imagine a scenario. First of all, a couple of things: the seroprevalence rates in Queens and in Staten Island after the epidemic were quite low. Queens was higher at 2.6% and in Staten Island, less than 1%. So you’d have to imagine a scenario where a person was, say, subclinically infected in the year 2000 in Staten Island and in the next year was unlucky enough to get encephalitis from some other cause and then you test them for West Nile antibody and find IgM antibody. Well, we would not expect that outcome; that would be a very rare event. Second of all, we would not expect that the IgM would persist in the cerebrospinal fluid. Nobody’s going to do those studies because they’d be unethical. But at that point when that patient comes into your laboratory like that this summer, you’re going to test CSF for one thing and that will help you, but also acute and convalescent serum for development of IgG. So, you’ve got to look at the whole serological picture in that case. There should be no serious problem in sorting that kind of case out clinically. I think where it has more application or it causes more confusion is in seroprevalence studies where you might have only 1% of people who were infected the year before, and a smaller percentage of those are going to still have antibody later. So, I think it’s more of a theoretical concern than a practical impediment. Bob?

SHOPE: From what I’ve heard this morning, I think there might be a possibility that IgM in the spinal fluid might persist. I have no basis other than seeing some of those pathology pictures.

CAMPBELL: Well, I guess it’s theoretically possible, but from what I’ve heard, persistence of IgM and the CSF would imply that there is active plasma cell production due to ongoing infection of the cerebrospinal fluid. So, from what you saying—from everything we’ve been able to find in the literature, and Tony Marfin has done a lot of the literature review in this—there’s not that much information available. But in the human IgM and CSF it’s cleared very quickly, and unless you have active plasma cell production—unless there’s a plasma cell that’s being stimulated in that sequestered space—the IgM is cleared very quickly. So, that would imply persistent infection to most infectious disease physicians, I think. Maybe you’re right. But we’ll have to see. I don’t know how we’re ever going to figure that out, though, because nobody’s going to do serial lumbar punctures on patients.

SHOPE: Changing the subject a little bit, I meant to mention that Dr. Hall is here from Australia, and he’s developed a test for use in sentinel chickens, which might be more broadly applicable—I don’t know. But it might be worth hearing comments from him as to what we really need is a specific serological test, and his approach might be something that’s helpful.

CAMPBELL: At this point, we’d invite questions or comments from the audience.

LANCIOTTI: I want to make a quick comment about neutralization, and this is to build on what others, especially Dr. Shope, have had to say. In our lab, when I mention confirmation of the ELISA test by neutralization, that doesn’t mean we do neutralization just with West Nile. It’s important to do the whole panel of the flaviviruses that occur in North America. So last year we routinely did dengue, because of potential for travel, West Nile, SLE, and Powassan. In most cases when we can demonstrate specific sera conversion to West Nile, the titers to West Nile may have been 1 :512 or greater. There was an accompanying dengue and SLE neutralizing antibody titer much less—1:20 or 1 :40—but I would say in most cases we do see a cross-reactivity, and so I think it’s very important when you do the neutralization tests to include the other flaviviruses.

SHOPE: I think it may be important also to test dengue patients, especially the other way around using that same panel against convalescent sera from a dengue patient. And you may find that some of these crosses are more one-way.

DALE L. MORSE (New York State Department of Health, Albany, NY): A number of panel members have mentioned the need for caution in interpreting human serological tests and the need for quality control in confirmatory testing. It wasn’t well publicized, but a small number of the laboratories in New York State, New York City, and the CDC last year that had experience with this had some anxious moments as we followed up on screening tests on the serologic IgM that were low level titer or intermediate and as we tracked back to find out whether these really were going to be cases or not. Now as more laboratories are going to come on board, private and commercial, the potential testing that we’ll presumably be doing will probably be just a screening test because they don’t have BSL-3 or don’t have the capability to do neutralization tests. There’s a potential for a much larger concern and for following up all these tests or potential reporting of a potential case without confirmatory testing. New York has fairly strong regulations so we can try to have some quality control for those labs coming on, but a number of states don’t have that ability to regulate, so I have three questions in this regard. What are the capabilities nationally for quality control of commercial labs that might come on during testing in the coming year? Number two: What is the capability for confirmatory testing for those labs that are going to start showing some potential positive tests on the screening test? And third: Should there be a requirement for a two-step testing before labs can be certified to test for West Nile?

LANCIOTTI: We strongly recommend confirmation by neutralization. Last year, there were a handful of specimens that we couldn’t confirm by neutralization. And we’re following up on some of those to see what they could have been. But clearly, you need to have a confirmatory test. We would caution any state from reporting a result without confirmation. Now, if they don’t have the ability to do that, that is one of the things that CDC certainly is offering. Many states made use of this where they got a positive IgM, ELISA result, and forwarded the specimens to us for confirmation. If you can’t do that in your lab, I think you’d want to have the neutralization tests.

CAMPBELL: I want to ask when MRL [Materials Research Laboratory] may be going to release their test kit for general commercial use.

QUESTION: Well, in theory, they can’t because it’s not FDA approved, so they could only use a home brew at MRL. So they can’t even pass it off as specialty.

LANCIOTTI: We are trying to establish a relationship with MRL so that when they receive specimens—and they have done this—they’ll contact us for confirmatory testing. There’s also a need for improved communication that’s being set up between MRL and the states where the samples came from, because in some cases state health departments are bypassed, where samples are sent directly to MRL and the state epidemiologist is not even aware of it. So, we are trying as best we can to set up all that communication, but certainly MRL is aware, and we have received samples from them for confirmatory testing.

GRADY: I believe New York State also is going to require any laboratory that is permitted to do West Nile testing for New York residents to provide material for confirmation, either at the state laboratory or at CDC, Fort Collins.

ROEHRIG: I think that’s a good point. I mean, I think it’s kind of a double-edged sword here. It’s not necessarily bad to have commercial testing for West Nile because you’re casting a broader net to try to find these cases, but somehow the link has to be made for those laboratories to get confirmatory testing. I’m not sure the two-tier testing requirement is going to be the answer because if you do that, it’s going to just close out those laboratories that will never do neutralization. The important part is to get the confirmatory testing done somewhere and to make the link back to the state or back to us at the CDC. We can probably do the confirmatory testing. I don’t think that’s an issue.

[Dr. Duane J. Gubler (born 1939)] (Centers for Disease Control and Prevention, Fort Collins, CO): New York is probably one of the few states around that has a system in place where they do actual proficiency testing of local laboratories before they can be certified. And, in fact, do we have a representative here from APHL [Association of Public Health Laboratories]? APHL is really the key to the answer. We went through this with Lyme disease testing, and CDC has no regulatory authority to enforce this kind of proficiency testing. It’s really at the state level. We need to bring APHL into this discussion and develop a proficiency-testing scheme probably modeled after the New York program that certifies local laboratories that were testing. And, I agree with John that commercial tests are a good thing as long as they’re controlled reasonably well so that there is some quality control. Another solution is to use regional laboratories to do the confirmatory testing. The CDC can’t do it all, and New York can’t do it all, so we really need to work with APHL to set up regional laboratories to do the quality control confirmatory testing.

THOMAS P. MONATH (Acambis Inc., Cambridge, MA): The requirement to use BL-3 for West Nile and St. Louis and other viruses you’d like to use in performing neutralization tests is an issue for many labs. I’d like to gauge the interest of this group, in particular, on the question of interest in the way of converting these viruses to BL-2 agents so that one could use them safely for neutralization virtually in a large number of labs. Now, the approach that we suggested is to use our chimeric yellow fever technology, which is being applied for attenuated vaccines, so that one would wind up with a family of BL-2 viruses, or basically yellow fever 17D vaccine. But the envelope genes of West Nile, St. Louis, Japanese encephalitis could be used then because they’re specifically neutralized only by antibody against each of the respective flaviviruses. They could be used in the neutralization tests to distinguish sera that neutralized a specific virus of interest. So one could actually perform confirmatory neutralization tests in any BL-2 lab with a series of these chimeric flaviviruses that were specific for each of the virus types. Now we already have Japanese encephalitis and West Nile available, and St. Louis needs to be made, and we’re willing to undertake that. We’re actually working with Dr. Gubler’s group now. But it would be useful to know whether this, in fact—if we’re going the right direction with this— would be useful. Would it broaden the capability of labs to do confirmatory diagnosis? It might be very useful in the situation of surveillance of birds and other animals as well. It would be interesting to hear some comment from the panel or anybody else in the room to get a sense of how hard to push this, because it would take quite a bit of work to actually validate the technology.

ROEHRIG: For example, with Japanese encephalitis chimera, when you have it approved for human use as a vaccine, clearly at that point that chimera can be classified as a BSL-2 agent. How much characterization do you need to have as far as pathogenesis or virulence in humans before it really will go from a BSL-3 to a BSL-2?

MONATH: Well, we had classified the Japanese encephalitis construct to BL-2 long before human trials.

ROEHRIG: Yes, but that was a vaccine, vaccine chimera.

MONATH: Right.

ROEHRIG: But for West Nile, you’re talking about taking a wild type and trying to attenuate.

MONATH: Well, we have mutants now too. They’re even more attenuated than the wild-type construct. But, I think the answer is we need data, and there are very good animal models, both mouse and monkey, that can be used. The critical question: Is the chimera less neurovirulent than yellow fever 17D, which is a BL-2 human vaccine? We’re testing the appropriate animal system. I think in the case of West Nile, we’ll have those data, including monkey data, within a few months. And so at that point what we would do is currently BL-3; we would ask our biosafety, institutional biosafety committee and other collaborators to reevaluate this at BL-2. We might ask the subcommittee, Arbovirus Laboratory Safety, ACAB, to look at this question independently— or the CDC for that matter. But I think it’s just a matter of providing the data.

ROEHRIG: So you think preclinical data in monkeys and mice will be enough to downgrade it?

MONATH: Yes.

SHOPE: With Japanese encephalitis virus, is it necessary to use the chimera? The parent of the chimera has been tested in literally millions of humans in China.

MONATH: Right. You could use a vaccine strain.

SHOPE: Yes, the SA14-14-2 strain.

MONATH: We don’t have one for West Nile though. There are attenuated strains of St. Louis and you can make the same argument. But I think that the ideal thing is that the yellow fever backbone allows you then to use the yellow fever vaccine for comparison in assessing relative virulence. And so you have a very convincing story about attenuation. So I think for St. Louis and West Nile, that’s going to be essential.

SHOPE: I think you make a good case, yes.

ROY A. HALL (University of Queensland, Brisbane, Queensland, Australia): I want to respond to Bob Shope’s mention of the competitive assay that I briefly alluded to in my talk yesterday. Basically we designed it for animal surveillance that works quite well for primary infections. We have a batch of sera that CDC had sent to us from New York, and in the initial trial, samples came out positive in the ELISA—not all, but some. And we’ve yet to compare the data with neutralization tests, and Ann Hunt said she’s got the data and we’re going to compare them. But at this stage it’s still a bit early. It’s certainly worthwhile to evaluate it with West Nile.

NANCY HALPERN (New Jersey Department of Agriculture, Trenton, NJ): I have a question for Dr. Shope. You mentioned, I believe, a Japanese encephalitis vaccine for horses. I was wondering if you or anyone in the room might know the mechanism to allow its use while we’re waiting for the vaccine’s development in this country.

SHOPE: I don’t know the proper mechanism, but a great deal is known about the vaccine itself.

HALPERN: I know. I know it’s in use for horses, right.

SHOPE: It’s also been tested for its ability to infect mosquitoes. Dr. Barry Beatty and Dr. Tenbo Chong did those experiments; they’re published. It does not readily infect mosquitoes; since it’s alive, the attenuated strains are one of the things you’d be worried about. Maybe someone else knows how veterinary vaccines are licensed.

ALEXANDER E. PLATONOV (Central Research Institute of Epidemiology, Moscow, Russia): What use are the specific ligands for West Nile, for example—not like a physical concentration, but a specific concentration of a sample.

LANCIOTTI: I’m not aware of anyone who has done that with West Nile. So you’re saying concentration before one of the nucleic acid tests?

PLATONOV: Yes, for example, or before RT-PCR, before application of any diagnostic tools.

LANCIOTTI: Again, I’m not aware of anyone who has attempted that. I will say that when we use PCR or NASBA or any of those assays looking for West Nile in bird tissues, there is such a large amount of virus that in most cases it’s not really a sensitivity issue. There are always going to be a handful of cases where you would like to concentrate before, and mosquito pools are pretty much the same. There’s a large amount of virus, so that there are not too many ambiguous samples. It’s really in human diagnostics that we run into, I think, an issue either of sensitivity of the test or as likely the timing of the specimen. We’re looking at CSF or serum beyond the febrile phase. So there may not be any virus there at all, because we also didn’t detect virus by isolation as well.

GUBLER: I want to come back to your point about persistence of IgM antibody in the cerebrospinal fluid, Roy. I know I’ve read this in Kuno’s review of persistence on page nine—I just looked it up. On page nine he says—this is a reference paper published in 1983—monkeys inoculated with West Nile virus presented an asymptomatic infection, but the virus persisted in the brain for up to 5.5 months without viremia. So, in other words, they isolated virus five and a half months after, and so this would explain the persistence of the IgM if this is occurring in humans.

CAMPBELL: I don’t remember.

GUBLER: We need to keep an open mind.

CAMPBELL: Did they test cerebral spinal fluid for IgM?

GUBLER: Yes, well they didn’t test CSF, or at least he doesn’t cover it in here, but it was brain. But indicating that there’s virus in the brain meant it may be intrathecal production of antibodies.

ROEHRIG: Well, Duane, I don’t think that’s a problem, but I think that you’re comparing apples to oranges even though monkeys are closer to humans than mice.

GUBLER: I think the point is that we should keep an open mind.

ROEHRIG: Well, I think we should keep an open mind about everything. I don’t think that that’s an issue. But the point of having cerebrospinal fluid taps, to actually answer the question, is going to be hard to get for humans.

DENISE MARTIN (Centers for Disease Control and Prevention, Fort Collins, CO): I’d like to bring up another point that hasn’t been really driven home yet about confirmatory testing. There was a fair amount of concern last year about the length of time it takes to do a confirmatory test in several capacities. There was a lot—especially in states that were looking at the possibility of the first case in their state—of push for doing a neutralization on a very acute specimen last year. And a lot of these extremely acute specimens have not allowed for time for neutralizing antibody to be formed. I think we need to make the point that you can detect IgM by an ELISA test before you can detect IgM in a neutralization test. So, it’s extremely important to do a classical neutralization test and see a rise and fall in antibody to get that convalescent specimen. In the push to identify West Nile in each new state, often it doesn’t seem to be acceptable to wait for that convalescent specimen, and yet ruling out West Nile on a very acute specimen might be erroneous in the long run.

SHOPE: I think that one could make a case for having a diagnosis that’s called a presumptive diagnosis: that’s what you do I think. I guess I would be even more conservative, and even after you’ve done the neutralization test and confirmed the positive reaction that you still call it a presumptive diagnosis, unless you’ve actually identified the virus either by isolation or by sequencing, because of the cross-reactions that can occur in flaviviruses. I don’t know, is anybody else with me on that wavelength? Probably not.

ROEHRIG: I think I missed that.

SHOPE: I don’t think we ever have a definitive diagnosis on the basis of serology, and we should keep that in the back of our minds. I understand that the patient and the press demand that you tell them what they have, but we should always, as scientists, keep a little bit of doubt when we’re basing our diagnosis solely on serology. That’s a personal feeling.

ROEHRIG: So you would say if you had a patient that had presented with clinical encephalitis with muscle weakness, had a huge IgM response in the CSF, and in acute and convalescent specimens showed an 8- to 16-fold titer rise against West Nile encephalitis, and stable or no antibody against the rest of the cross-test as flaviviruses— this isn’t a dead person, so you don’t have autopsy specimens, and you can’t get PCR or isolation positive—that you wouldn’t be comfortable that that was a West Nile case?

SHOPE: I would be comfortable on a presumptive basis (general laughter). I think one of the keys to your statement is that you’ve tested it against all of the other flaviviruses. We continue to find new flaviviruses. In fact, New York City is a case in point; one did find a new flavivirus.

ROEHRIG: Yes, and as we did the neutralization tests, it became clear that it wasn’t St. Louis (general laughter).

CAMPBELL: I want to say one thing: Earlier, Rob Lanciotti used the term “gold standard” in terms of serology, and I think, just to go back to basic principles, most of us would agree that if you’re really a purist that the gold standard for the diagnosis of West Nile virus infection is isolation of the virus from the patient, and that’s rarely possible. So we all realize that serology is not the same as viral isolation.

GRADY: I should throw in too that we had one case in New York that, based on the capture IgM, was suspect for West Nile, and it turned out to be dengue when the PCR was done on serum. When the folks followed it up, it turned out that it was a travel history; the patient had actually become ill while in Haiti, but didn’t bother to seek medical attention until back in New York.

PATRICIA REPIK (Charles River Tektagen, Malvern, PA): This question is for Rob. Are you recommending a virus isolation by any standardized method, say, looking at it for over a period of however many days and doing subpasses, like one subpass or two subpasses? Or is it just plaque isolation? Are you allowing replication to proceed?

LANCIOTTI: There are two ways we do it. For the mosquito testing, we actually just do it with an overlay and look for plaques. But in the human, trying to isolate virus from the human cases, we went the extra mile and did multiple tests and liquid culture and multiple cell lines, and mice, as well. So, we tried everything that we had available to us besides direct mosquito inoculation. I don’t think we ever tried that. So, for humans I think that’s what we would recommend just because we try everything that is available, not just mammalain cells, but C6-36 cells as well.

ROEHRIG: There’s a practical side to this. I mean, we can do that as a reference laboratory, but in the heat of trying to diagnose outbreak at a state laboratory, a local laboratory gets to be a bit tedious when you have a thousand other samples coming in and trying to deal with them. Duane, didn’t Bruce do mosquito inoculation with some of those West Nile specimens?

GUBLER: No, unfortunately we didn’t do mosquito inoculation, but I think Bruce did show that the fluid culture of Vero cells was probably the most sensitive of the systems. They were all fairly close. Laura and I were just talking about this yesterday. We should do mosquito inoculation in some of these samples. We haven’t done them.

HELENE PAXTON: (PanBio InDx, Inc., Baltimore, MD): Rob, at the ASTMH meeting when we had our first West Nile meeting and Duane was chairing it, you talked a lot about using IFA and using slides that had all of the relevant viruses in the United States or outside the United States on your slides so that you could do cross reactivity studies. In the last two meetings, there’s been very little said about the usefulness of indirect IFA. As a commercial company, we constantly get barraged with whether you’re going to do an IFA titer. Is there any strength in doing, say, a 1:40 screen and a 1:320 screen for each of the viruses that might be circulating in the United States? Can you define any true seropositivity? From what Bob said it sounds like you probably can’t, but we do get asked that question an awful lot. The other issue is whether or not, in animal models, if you’re doing chickens or horses or some of the other mammals, whether or not the IFA would hold up better. Because most labs—and I’m getting ten or twelve phone calls a day—can’t do the more sophisticated techniques, and they want to have a way of screening to see whether or not there’s a possibility that West Nile is in the differential diagnosis.

KRAMER: I can answer for use of IFA for bird serology. Susan Wong can answer for the human serology, but we do indirect IFAs on all our ELISA-positive bird sera as the second test, and then if that still is positive we go on and do a plaque-reduction neutralization assay. And it’s been very useful in cutting back the number of neutralization tests we need to do, so it’s one more refined step after the ELISA. We also use an indirect IFA to confirm positivity in tissues and in cell culture after the cells are inoculated. If cytopathology is seen we spot slides with the infected cells and do IFA on that, and not just for West Nile but for other viruses that are active in New York State.

PAXTON: That’s direct FA though, right?

KRAMER: No. Indirect.

PAXTON: You’re doing an indirect on the tissues?

KRAMER: Yes. Those were the slides that I showed yesterday and that were in Alan Dupuis’ poster.

PAXTON: So you’re using positive sera against a substrain as your antigen that could possibly be infected?

KRAMER: We’re using the antibody to make sure that what we’re seeing is the virus.

PAXTON: I have one other comment. Duane knows this well: the issue of getting FDA clearance. I talked to Woody DuBois at length about this in the microbiology branch, and he is very aware that there is a tremendous need to get these tests cleared. You can’t purchase IFA slides as analyte-specific reagents from various manufacturers, and the analyte-specific reagents statute requires that you make a disclaimer in your patient reports saying that this has not been cleared by the FDA. Whole kits have not been cleared and are more problematic, and so you may have a situation where you’ll find components floating around where their “label” is ASR. That’s really not the intent of the law. But Woody has strongly suggested that if you could get dengue cleared, then West Nile could be treated as a predicate against dengue. If the major problem of West Nile is getting enough samples to be able to do an FDA submission that would hold up, then let’s keep that in the back of our minds as we’re collecting samples, because a commercial company totally has their hands tied behind their back in trying to get a clearance for an assay without having the proper samples.

ROEHRIG: I think that’s a good point. I forget these things, but positive control specimens are really difficult to come by, and luckily over the last couple of years we’ve had cooperative physicians and cooperative patients, like Dr. Asnis pointed out in her talk, that have been able to donate serum specimens for use as positive control. So for the physicians around, keep this in mind, because this’ll be an ongoing need, and you’d be surprised how quickly you go through it. This doesn’t include the needs of a commercial company where they need large amounts of these specimens banked.

SHOPE: Let me make a comment. We found that dengue convalescent sera worked beautifully in the capture IgM ELISA for West Nile, and you can get large quantities of that. I don’t know whether that would be permitted as a positive control by the

FDA, but they work.

PAXTON: That’s the problem, that the dengue is so cross-reactive.

SHOPE: You need to market the test as specific; your test could give you a flavivirus diagnosis, and that would be very useful I would think.

CAMPBELL: Dr. Bresnitz?

EDDY BRESNITZ (New Jersey Department of Health and Senior Services, Trenton, NJ): In the clinical setting where you have a patient who has the right clinical syndrome for encephalitis, and treatable infection has essentially been eliminated, nailing down the diagnosis, when you feel where it’s no longer suspect or probable, really has no either clinical or public health practicality to it, to be honest with you. From a clinical perspective, once you nail down the diagnosis it really doesn’t make a difference by then. And certainly from the control of mosquitoes it makes no difference either because it takes so many weeks. And I would argue that, in fact, it’s what we will do in New Jersey as well: it’s letting the mosquito control people know when you have an individual you suspect might have West Nile virus encephalitis—particularly if you’re coming from an area where activity has been demonstrated in birds or mosquitoes or horses. That waiting for diagnosis, final diagnosis, or even a probable diagnosis is too late for enhancing mosquito control. So it’s a nice theoretical discussion. From a practical perspective it really doesn’t make a difference.

LANCIOTTI: We would agree that waiting is probably better, but I can’t tell you how many times this year we’ve gotten calls that said we want you to confirm our ELISA result, and we have a press conference tonight at six. So, that’s not uncommon.

SCOTT P. HENNIGAN (Massachusetts State Laboratories, Boston, MA): Is anybody looking into using the Western blot? We use the Western blot for HIV infection; we use it for Lyme in a two-tier method. Is it something that could be used as a quick three-hour—I don’t like saying confirmation—but that type test. Instead of waiting for a couple of days for the neutralization, you can actually run a Western blot; you could probably do an IgM or IgG on it also.

ROEHRIG: I come from a laboratory that, before West Nile, dealt with Lyme disease. Where the Lyme disease test is, the Western blot is used as a confirmatory test for Lyme disease, which is, of course, a bacterial pathogen that has multiple antigens. Part of the differential for the Lyme disease Western blot is how many of these antigens are positive, how many are negative, and how many bands are seen. Flaviviruses are fairly simple organisms. There are really three proteins that you can see on a gel, one of which is the M protein, which is a very small membrane-bound protein that is immunogenic, but not overly immunogenic, so you’re really looking at two bands on the blot that may actually elicit antibody responses. The nucleocapsid protein, for all viruses and including flaviviruses, is relatively conserved, and so you would get very little serospecificity out of a reactivity with a nucleocapsid. That also needs to be factored in with the nucleocapsid because it interacts with the nucleic acid. It’s a highly charged protein and typically gives a very high background in a Western blot. So really you’re talking about the envelope glycoprotein as being the most important antigen, and since you’re really talking about one band, you can’t separate out five or six bands’ reactivity patterns. When you talk about the glycoprotein you’re talking about other problems with the glycoprotein, which is the confirmational stability of epitopes, whether you run Western blots with mercaptomethanol sulphides or whether you leave it native. The problem is that the protein really isn’t big enough to do a lot of differentiation from, let’s say, a dengue serum from a West Nile serum from an SLE serum. So I don’t think that the Western blot is going to help improve the specificity of the test. We’ve run a lot of Western blots over the years in my laboratory, and basically this is what you see. I could be wrong, and I want you to keep an open mind, but my best guess would be that a Western blot would not be specific for West Nile infection. A caveat could be that there are also nonstructural proteins that are synthesized by these viruses. A number of years ago, even before I came to CDC, Fort Collins, Dr. Dennis Trent and Ateef Qureshi looked at the immune response as five proteins, and their preliminary evidence was that this was a far more specific protein than the E glycoprotein. The problem with the NS5 protein is that not everybody makes antibody to NS5 to high enough levels that you can detect it on a Western blot. So if you did that sort of thing, you couldn’t use virion E (protein); you’d probably have to use infected cells where you’d be looking at the nonstructural or structural proteins. I think that it’s a thing that could be used, but we would have to do a lot more work with the specific virus to try to determine whether or not it would be useful. My initial take on it would be that it’s probably not going to add much over what we already have.

MONATH: I think the future is going to be very interesting with respect to diagnosis, because, fortunately, so far, West Nile has invaded an area of the United States where other flaviviruses are not so active. But as and if it moves west, West Nile will infect people who previously have seen St. Louis encephalitis. In fact, in the Midwest, there were big epidemics in the mid-1970s. This is the elderly population now that is going to get sick, so the complexity of flavivirus serology, serological diagnosis, is really highest where you have multiple flaviviruses circulating and affecting human population. And the complexities of the patterns of antibodies with respect to cross-reactivity are going to really be problems when this virus appears in the St. Louis country. So I think—if you want to look into the future and prepare yourself for it—that’s how that is going to affect our ability to make a specific serological diagnosis. It would be a worthwhile model. We really don’t understand very well what happens in these sequential infections to see what happens in the St. Louis–infected animal that’s later infected with West Nile. It’s going to be a real headache when that happens.

ROEHRIG: Well, at least, Tom, from a public health perspective, if we diagnose it as a flaviviral encephalitis, they can use control and, we hope, try to get it under containment. But yes, actually trying to define exactly what the agent is could be problematic. It will be problematic.

CAMPBELL: It reminds me of something that Dr. Bill Reeves from the University of California-Berkeley reminded Duane and me of recently. Tom, you may know this, but there was some early work—and again, going back to John’s point about looking at the early work and trying to appreciate what was already done and not reinventing the wheel; there was some laboratory work done by some scientists years ago, including Dr. Hammon, who is a pioneer in arbovirology—looking at cross-protection of dengue antibodies against West Nile virus in laboratory animals. And there is some cross-protectivity there, and I wonder what we know. I’m not familiar with any similar studies with SLE, but I would suspect there is some cross-protection at least.

SHOPE: There’s a large literature on this. Dr. Winston Price did a whole series of experiments in primates. And, yes, there is cross-protection as Tom says.

CAMPBELL: It could be that those people that were infected in Chicago in 1975 are going to be the least likely to get West Nile encephalitis now that they’re old.

GRADY: Before we end our panel, I’d like to throw out something else for your consideration: Although West Nile is new, it hasn’t replaced the usual causes of viral encephalitis that one sees, particularly enteroviruses and herpes group viruses. And you’ve got to keep in mind that that has to be part of the differential diagnosis. We have data suggesting that in the same time period in the greater metropolitan and surrounding counties, that there were probably, at a minimum, as many cases due to other viruses as there were due to West Nile. And I say at a minimum because enteroviruses can be recovered by culture, and there are a substantial number of laboratories that have that capability. We don’t know what their results were, and likewise more and more labs are coming on line able to do PCR for herpes group viruses. Again, we don’t know what their results are. We only know the results of specimens that were tested in our laboratory. Before we leave, I’d like you to remind you to keep that in mind. The outbreak doesn’t replace other viruses; it merely is superimposed on what is normally going on.

KRAMER: I wanted to add to that—Dr. Monath reminded me. I mentioned that we do cell culture because we don’t want to miss other viruses that are there. And I think it’s very important that as this virus moves into Florida, where there’s St. Louis, and as it moves toward the West Coast where there’s St. Louis, that if people use PCR to look for West Nile in mosquito pools, they might be missing St. Louis. So it’s very important to do assays so that you don’t miss the other viruses that are there on the mosquitoes as well, because you may not see human cases that will tell you there’s another virus there, except that they may be spotty. We don’t want to narrow down our focus on West Nile alone, even in the mosquitoes.

CAMPBELL: OK, it’s time to wrap it up. Thank you very much to our excellent= panel.