Advertisement Body PRO Covers CROI 2009, February 8-11, 2009

Rapid HIV Antibody Testing May Have Limitations in a Population With High Incidence of HIV

An Interview With Joanne Stekler, M.D.

February 11, 2009

My name is Dr. Joanne Stekler. I'm an acting assistant professor at the University of Washington. I also work for Public Health Seattle and King County. This is a poster entitled "Limitations of HIV Antibody Testing in a Population with a High Incidence of HIV Infection."1 Essentially, this is a summary of results of our HIV testing program for the last five years.

We realized after North Carolina started doing pooled nucleic acid amplification testing, or NAT, that we would start doing this.

In September of 2003, we became the second program in the U.S. to initiate pooled NAT testing. We're doing it only for men who have sex with men [MSM], because our population in Seattle is really focused in men who have sex with men. This is a population with a prevalence of about 15 percent and an incidence of .7 percent, so .7 per 100 person-years.

These are people seeking HIV testing through Public Health Seattle & King County funded sites. It was initially started in the Public Health Seattle & King County STD Clinic, and then has expanded to Gay City Health Project Wellness Center, and to our three Seattle bathhouses.

Those are the results of our analysis. We recommend that MSM in Seattle, if they are high risk, and high risk is men who report unprotected anal intercourse with partners who are either HIV positive or of unknown sero status, who report a bacterial STD [sexually transmitted disease], or methamphetamine or popper use in the last year -- that they get quarterly testing, so testing every three months. If they are otherwise low-risk MSM, we recommend that they get tested annually.

These are just the records from our electronic medical record. Of course, because this is a public health program, and not our actual research study, we have used a variety of different tests over the years.

When we initially started the program, our basic antibody test was the first generation -- Vironostika. In 2006 we changed to the second-generation assay, estimated to have a window period of about one month.

The rapid HIV test that we have used has been a combination of oral fluid and finger-stick testing -- depending on the site and depending on the year. So things have changed back and forth. It's a combination of both data, and I won't be able to give you separate data for that.

As I said, in 2003 we started doing nucleic acid testing. It was initially with an assay that was provided to us by Gen-Probe in pools of 30. People would get screened for HIV antibodies and then -- if they were antibody negative -- their specimens would get pooled together into three intermediate pools of ten samples. Then those three intermediate pools would get pooled into one master pool.

That was the way we did it from 2003 to 2005. At the end of 2005 we identified a person who was rapid tested who was actually found to be antibody positive by a first-generation assay, that Vironostika assay. (The rapid test we used was an OraQuick test, made by OraSure.)

So since 2005, if men have a negative rapid test, we actually screen them with our standard EIA [ELISA] before putting them into the pools. We do that in order to get the results back quicker. Because we can get EIA results back in one or two days, as opposed to the nucleic acid testing, which takes about one week.

I'll talk now about the results. Basically, we have two populations of men who have sex with men: those who initially tested by the standard blood test, and then those who initially tested by rapid test.

Of the 7,041 people who tested initially by standard antibody testing -- again, these are predominantly MSM at the STD clinic -- 159, or 2.0 percent tested positive by the early generation EIA. An additional 13 men, or 0.2 percent, were EIA negative and NAT positive.

We are reporting sensitivity, because that's different from the way they get F.D.A.-approved for sensitivity. EIA picked up 159, of 173 HIV infected persons who were testing through the program. NAT picked up 172 out of the 173. One person actually tested NAT negative, four days after acquiring HIV, giving us sensitivity of 92 percent for EIA, or the antibody test, and 99 percent for NAT.

Among the 6,964 people who were rapid tested, 153 people, or 2.2 percent, were OraQuick positive -- again, a combination of oral fluid and finger-stick sampling. Another 16 people were OraQuick negative and EIA positive, and most of them have early infection. That's 0.2 of the testing population. Another 23 people were EIA negative, NAT positive.

That comes to a cumulative sensitivity of 80 percent for OraQuick, 88 percent for the algorithm of OraQuick followed by EAI; and then, all of the infections that we know of were picked up by NAT.

We were able to get people their results at a median of 19 days -- if they were anonymous testers -- and 16 days, if they were confidential testers. And there's no difference between those.

The next part of the poster [click here to view poster] represents the 16 people who were OraQuick negative, EIA positive. It shows that they were about equal distribution, tested initially by finger-stick and oral fluids, that the first two were by -- the first generation EIA that were remaining, or the second generation -- and then their western blot results.

What you see is that all of these folks who tested OraQuick negative: 14 of the 16 actually had positive western blots. The other two: one of them was an early infection and then seroconverting. The other person probably had late-stage infection, with a very low CD4, and never progressed in his western blot.

The next part [click here to view poster] represents a sampling of frozen specimens that were performed by Abbott in order to identify the relative sensitivity of the fourth generation combination assay, their ARCHITECT assay, compared to NAT.

We sent them 16 specimens from persons who were EIA negative, NAT positive. Their combination assay was able to identify 15 of the 16. The one that was not identified had an HIV level of -- depending on the assay tested -- about 10,000 copies per mL. Overall, the ARCHITECT combination assay was able to pick up 15 of 16, 94 percent, of our acute infections.

Those are really the main results. We think it's really important that, in populations at high risk, with high incidence -- like the men who have sex with men in Seattle, that we use the best tests possible.

Our findings may not be generalizable to other populations that have lower HIV prevalence, or lower HIV incidence, or don't test quite as frequently as MSM in Seattle do.

Is the NAT available throughout the country?

The NAT is basically a viral load test. There is one NAT test that is made by GenProbe that is F.D.A. approved for the diagnosis of HIV infection.

The remaining of the assays, including ours, are home brewed, quantitative, real-time, RT, PCR amplification assay that we're using off-label, in order to diagnose acute HIV infection. But it is available, and there are potentially some commercial labs that can make this available to people if they want them.

Has this study been corroborated by other similar studies?

The results that you're going to get vary a little bit, depending on the population that you test. What we look at is the increased yield of NAT over antibody testing, so that, in many of the populations tested, you get on average about 10 percent increased yield. For every 100 antibody tests that you have positive, you'll get 10 people who are antibody negative in nucleic acid test, and amplification test positive.

Again, it varies. There are some places that have had higher yields, and there are a fair number of places that have had lower yields. It's very dependent on the population, the incidence, and the likelihood that they are testing in the window period.

What is the Department of Health in Seattle doing now? Are they going to be using this from now on?

We have been using this. We consider this to be standard care in Seattle, for folks who have access to it. We have expanded our pooled program to some of the community providers who see men who have sex with men, in order to increase the diagnosis.

Part of the problem is that, yes, we have diagnosed close to 40 cases -- actually, at this point, over 40 cases -- of people who are EIA negative, NAT positive. But over this time, more than 1,000 people were infected by HIV. So we're diagnosing a very small percentage of all people who acquire HIV.

We need to increase the diagnosis of acute HIV -- because folks who are more infectious are more likely to transmit to their partners. There also may or may not be some benefit to treating people with [antiretrovirals] during acute infection.

The ways to do that are either to increase the frequency that people at risk get tested, and maybe also to have people who have symptoms of acute infection come in for RNA testing at that time.

The symptoms include a flu-like illness. People experience a fever, rash, sore throat, stomach illness. It's very variable, depending on the person. But maybe up to half of the people will have some sort of symptoms related to HIV acquisition.

How does the cost compare to the other tests?

The addition of the nucleic acid testing to our established program costs, based on an analysis that I did a few years ago, just about $7,000 per case of HIV detected. That's comparable to, I believe, our bathhouse testing program. It's about that much, if not a little bit less.

That's not a true cost effectiveness analysis, but doing a true cost effectiveness analysis -- in terms of HIV cases avoided or sort of quality of life years - would require doing a lot of assumptions. But we feel that it's cost effective in our population, based on what are the other established programs. That cost effectiveness is going to depend, just like the effectiveness does, on what your population looks like.

Thank you very much for taking the time to summarize your study!

You're welcome.


  1. Stekler J, Swenson P, Coombs R, et al. Limitations of rapid HIV antibody testing in a population with high incidence of HIV infection. In: Program and abstracts of the 16th Conference on Retroviruses and Opportunistic Infections; February 8-11, 2009; Montréal, Canada. Abstract 990.
    View poster: Download PowerPoint

This article was provided by It is a part of the publication 16th Conference on Retroviruses and Opportunistic Infections.

Please note: Knowledge about HIV changes rapidly. Note the date of this summary's publication, and before treating patients or employing any therapies described in these materials, verify all information independently. If you are a patient, please consult a doctor or other medical professional before acting on any of the information presented in this summary. For a complete listing of our most recent conference coverage, click here.