February 9, 2009
My name's Chris Pilcher and I'm a HIV researcher from UCSF [University of California, San Francisco]. I'm principally interested in HIV diagnostics and particularly in detecting and managing acute HIV infection.
The poster is appropriately titled, "A Clinical Study of Antigen/Antibody Rapid Testing for Acute HIV Infection."1 This is a study that is unique in a couple of ways. It's essentially a head-to-head comparison of HIV diagnostic tests for their ability to detect cases in testing populations. Since 2003, the San Francisco Department of Health has been using HIV RNA testing in their testing algorithms to retest specimens that are tested for HIV by antibody tests and found to be negative.
Chris Pilcher, M.D.
A lot of early work, including work with this testing program in San Francisco, has indicated that particularly in high-risk testing populations, such as STD [sexually transmitted disease] clinics, and non-traditional testing venues, for instance community health centers in gay communities,-- that these acute infection actually can represent as many as 5% or 10% of all the cases that are out there. This raised the question, which we tried to answer in this poster: How do the variety of available and now newly available HIV tests perform given that there is so much acute infection that was previously unrecognized in these testing populations?
The populations that were studied in this were in about 20,000 testers in San Francisco, mostly since 2003, who participated in a whole variety of programs all targeted to high-risk folks in San Francisco, predominately men who have sex with men [MSM]. All of these patients had blood drawn, had a specimen sent for either a rapid test performed at the site or a specimen simply sent off to the lab, but all of them did a have a specimen sent off to the lab.
Then the central laboratory -- which is the public health laboratory in San Francisco -- did testing of all of the antibody negative specimens to look for HIV RNA and did testing of the antibody positive specimens to try and confirm the diagnosis.
When specimens were RNA positive, but negative or inconclusive or indeterminate on the standard antibody testing that the public health laboratory did, we then took all of those problem specimens -- the unresolved, probably likely acute, but not always acute infections -- we took all those specimens and we retested those against all the array of available and developmental assays that we could get our hands on to examine whether or not those tests could do better than the testing that was done at baseline.
Comparing the results of all the different assays as they performed both in the clinic and then against the panel of problem specimens in the lab, we've made a couple of important observations.
The first was that oral fluid testing -- using the OraQuick Advance device on a oral fluid swab -- detected only about 86% of all of the cases of HIV that were actually present at testing. And this is because it missed not only all of the acute HIV infections in that 14%, but also some antibody positive infections that turned out to be positive on standard -- other antibody tests and on the western blot. So we identified that the OraQuick had not performed very well when we used oral fluid.
Has this been seen before? Is this well known?
There have been many reports but there -- the contribution of acute infection to the problem, if you will, with oral fluid testing hasn't previously been appreciated, so that was an interesting finding.
We did find though that the wide range of immunoassays -- which are central laboratory tests as well as point-of-care rapid tests -- performed very similarly, except for some new investigational assays that actually perform very, very well in detecting acute infections.
The first is a very exciting new test, which was provided to us by the company that makes it -- Inverness. It's called the Determine Antigen/Antibody Combo test and it detects not only antibodies, but also HIV p24 antigen, which is the most common protein in the blood made by HIV and is detectable by a number of different tests, but this is the only rapid test that's capable of detecting p24 antigen in addition to antibody.
So it's a saliva test?
No, it's blood. And the Determine Combo assay that looks at antigen, as well as antibodies, detected about half of all of the missed infections by standard, other array of rapid tests, so that was important compared with the other blood tests which missed about 8% of infections. The Determine test only missed 4% of infections. So it picked up about half compared with standard, OraQuick and other rapid tests.
The last finding that was important was one of the new investigational central laboratory tests, which is a large central lab machine that does what's called an immunoassay, which is also designed to detect p24 antigen, but in a much more sensitive fashion, actually did terrifically well, missing only 2% of cases that were RNA positive. Those cases that were missed actually had very low viral loads, so they were in very early acute infection when the virus was still ramping up.
We concluded, on the basis of all of these results, that the Determine Combo test, which was novel appears to perform well in detecting some acute infections, but not all acute infections that are missed by other antibody tests. It does substantially increase case identification when compared with other antibody rapid tests that the new fourth generation assay, called the Architect Combo assay, for use in the central lab detects nearly all -- about more than 80% of the RNA positive acute infections that are missed by the standard test in the clinic.
Finally that the OraQuick Advance test -- while it works very well on blood -- when it's used for oral testing, it is significantly less sensitive. And I think that's an important thing to take home. Now I don't mean to imply that the OraQuick oral fluid test isn't a good test. In fact, I think it has potentially a very important role in a lot of testing settings. I just would emphasize that the use of oral fluid testing in very high-risk settings, such as these high-risk testing programs in San Francisco needs probably to be backed up by the use of some other more sensitive test in addition to make sure that patients don't have acute infection.
A couple of questions. How are these tests being used now? Are they approved for use and are being used or are they experimental?
It's different for the two tests. The Determine rapid test, the combo rapid test, is undergoing regulatory approval in Europe and has completed its initial sort of pre-market studies, but has not yet been introduced into the market. So it's available only for research use.
The Architect is actually one of many similar assays that have been approved by regulatory agencies in Europe and Australia and elsewhere. And they're now very widely used outside of the U.S., but none of the diagnostic companies has yet put in a claim with the F.D.A. [Food and Drug Administration], although it may be something that we can look forward to. So right now, both of those newer tests that detect acute infections are investigational research use only, but not for long.
What is the San Francisco Department of Health doing with the understanding of this finding, how are they changing practice?
This finding is about two weeks old, so nothing yet. They're going to be reviewing the data and I'd refer you actually to the Department of Health to talk about that.
How does your study -- there's a whole bunch of studies here that are showing, I guess, somewhat similar and somewhat diverse things. Can you talk about that a little bit?
A number of studies have also examined the ability of the Architect assay, in particular, to detect acute infections, but most of them have focused on the acute infection specimens and have not looked at the denominator of antibody positive infections that are out there to actually see what the impact of test choice would be on the actual effectiveness of the testing program.
In this analysis, we were able to capitalize on the data from large testing programs involving over 20,000 people where all of the patients actually were very certain that they either did or didn't have acute infection. So we got a nice picture of the numerators and denominators that spell out the differences between the tests. That's why it's unique and this is the first research study to my knowledge that's been presented here on the use of the Determine test, which is the first rapid test for acute infection. For those of us who are in HIV acute infection prevention business, having a rapid test is sort of the holy grail of making acute HIV prevention a really potentially effective strategy to reduce transmission.
Meaning treatment as prevention?
Actually, I think counseling as prevention can be effective in certain settings. And that's a whole other conversation.
Thank you very much.
Sure, thank you.