Abstract
During the 2010 HIV Diagnostics Conference, which took place in Orlando (FL, USA) between 24 and 26 March 2010, salient new data related to three aspects of HIV testing (rapid point-of-care testing, laboratory assays and new technologies) were presented and discussed. A conundrum central to HIV diagnostics in the USA for the last few years has resulted from technological developments in HIV testing that have outpaced recommendations for screening and confirmatory testing. Perhaps in response, one of the major outcomes of this 2010 meeting was the proposal of a novel laboratory testing algorithm. The proposed algorithm aims to take advantage of all of the capabilities of currently available tests (sensitivities for HIV-1 and HIV-2 IgG and IgM antibodies, and p24 antigen), while at the same time enhancing the turnaround time of results. This report will summarize the presentations at the meeting.
Rapid, point-of-care testing
It has been recognized since the inception of HIV testing that the capabilities of even the best laboratory-based tests are for nothing if infected individuals fail to get their test results and enter care. Rapid, point-of-care tests for HIV resolve this problem by providing HIV test results to patients within minutes of test execution. Such tests are simple and their complexity is deemed ‘waived’ by the US FDA; hence they can be performed by a broad range of trained personnel. This has considerably expanded testing by non-laboratorians and significantly amplifies the importance of pre- and post-test information. In an effort to guarantee that patients get accurate HIV test results, algorithms have been proposed involving secondary or even tertiary rapid testing after initially reactive rapid test results.
Public health departments from New York State, New Jersey, Los Angeles and San Francisco (USA) described evaluations of multirapid test algorithms. All sites agreed that between 35 and 50% of all individuals who test positive on an initial rapid test fail to receive their confirmatory test results when such confirmation is based upon laboratory testing (e.g., western blot). All four locations had instituted either two or three rapid test algorithms at point-of-care for evaluation purposes.
Data from New York and San Francisco indicated that a second rapid test was useful in the resolution of a false-positive initial rapid test. In San Francisco, 20% of initial rapid test-reactive specimens were negative on a second rapid test and later by western blot. Similarly, in New York, approximately 22% of reactive initial rapid tests were negative by both a second rapid test and by western blot at the laboratory. Results from similar testing volumes in New Jersey showed that only 7% of the initial reactive rapid-test results were discordant with the second rapid test; these were also negative at the laboratory by RNA and western blot. Comparable data were presented by the Los Angeles Office of AIDS Programs and Policy, where approximately 6% of initial reactive rapid tests proved to be false-positive. All four sites stated that two rapid test algorithms increased the degree of linkage of care for positive individuals, but only New York presented quantitative data in this regard (90% of positive clients confirmed to have entered care after testing positive by two tests at point-of-care compared with the 35–50% range reported by many other sites). Moreover, there was consensus among the four locations that counselors very much preferred having a second rapid test available. As reported by San Francisco, it helped to alleviate potential ‘psychic harm’ on the part of the client.
In response to these presentations, questions included: is the benefit of running two rapid tests worth the time and expense? Is the benefit of the second test more for the counselor or the patient? Does the order of specific rapid tests in a multirapid test algorithm matter? For example, some rapid tests might be more sensitive for recent infection than others. How might this affect discordance between the first and second tests? A key term arose that subsequently pervaded the conference: ‘orthogonality’, a concept whereby the tests that are suitable for use in combinations must employ different antigens or principles. Put another way, some rapid tests of different brands use recombinant antigens from the same supplier and are therefore essentially identical. This must be taken into account when designing multirapid test algorithms.
Laboratory testing
Within the realm of laboratory-based testing, the 2010 meeting covered numerous diverse topics. The largest number of presentations were related to recency testing and antigen–antibody combination assays (also known as ‘fourth-generation’ immunoassays). Other topics of interest for laboratory testing included the value of signal-to-cut-off ratios in interpreting enzyme immunoassay (EIA) results, the investigation of reported increased false-positivity among pregnant women and the use of tests other than western blot for confirmation. Below, summaries are provided in the order in which they were presented.
Data presented by the CDC illustrated the potential utility of immunoassay signal-to-cut-off values in diagnostic algorithms. Truly antibody-positive specimens tested by the same or different EIA generally generated high signal-to-cut-off values. Specimens that were false-positive on an initial EIA very frequently gave very low signal-to-cut-off scores when repeated on a different EIA.
CDC data also showed that the rate of false-positive results from contemporary EIAs was lower among pregnant women (rate: 0.14%; 921,438 tested) than among nonpregnant women or men (rate: 0.21%; 1,103,961 tested). However, pregnant women who tested repeatedly reactive by EIA were more likely than others to test negative or indeterminate on western blot. The presenter commented that HIV was less prevalent in pregnant women in the USA, but that pregnant women are more likely to possess immunoglobulin that cross-reacts with p24 antigen.
Data from the Blood Systems Research Institute (CA, USA) elegantly showed that, among specimens from approximately 3.6 million blood donors, substituting an indirect immunofluorescence assay for confirmation in lieu of western blot led to a more than tenfold reduction in indeterminate results.
In what proved to be a rather prescient presentation, the University of Washington (WA, USA) presented the results of an investigation of Bio-Rad Multispot HIV-1/HIV-2 Rapid Test (Bio-Rad Laboratories, WA, USA) after repeatedly reactive EIA screening. Use of this moderately complex rapid test for confirmation reduced turnaround time for results by 2 days, allowed detection of two occult HIV-2 infections (which would have been classified incorrectly as HIV-1 by the currently approved western blot) and yielded 100% specificity. This presentation was of particular interest because the meeting closed with a proposed laboratory testing algorithm that would call for a HIV-1/HIV-2 differentiation assay similar to the Multispot test as a potentially new choice for corroborating reactive screening tests.
Recency testing
The incidence of HIV infection (the number of new infections that occur per year) is often only determined with complex observational cohort studies. Laboratory techniques for estimating incidence generally rely upon modified versions of assays that have sometimes been withdrawn from the market shortly after their use was validated. Several presentations described novel methods for assay-based incidence estimation. Data presented showed that the Ortho VITROS® HIV 1+2 Assay (Ortho Clinical Diagnostics, NJ, USA), could be modified to generate an estimate of ‘recency’ either by a ‘detuned’ method (diluting the specimen 1/400 prior to testing) or by an avidity modification (comparing results of the standard assay with those from a specimen mixed with the chaotropic agent guanidine to disrupt weak antibody bonds). Based on seroconversion panels tested with these techniques, the test could define a mean recency period of approximately 8 or 6 months, respectively. A CDC group created an avidity modification of the Genetic Systems HIV 1/2 Plus O EIA (Bio-Rad Laboratories) by use of diethylamine as the chaotropic agent with similar results. Another CDC investigator presented intriguing data showing that HIV-specific IgG3 provides an excellent biomarker for discerning recent infection. In a bead-based Luminex® format directed against different HIV-1 antigens, an IgG3 assay defined an approximately 8-week window period, with the least variation and misclassification, raising the possibility that it might have clinical utility for defining recent infection in individuals in addition to applications for incidence estimation.
Antigen–antibody ‘fourth-generation’ immunoassays
In widespread use outside of the USA for several years, immunoassays that detect both antibody and viral antigen are only now being submitted for FDA approval. Three companies presented data regarding their fourth-generation immunoassay products: Abbott Diagnostics (IL, USA), Bio-Rad and Ortho Clinical Diagnostics. All three companies showed that, with seroconversion panels, such assays could yield a 10–12-day reduction in the window period relative to third-generation (IgM-sensitive) antibody assays, hence allowing for the detection of nearly 80% of acute HIV infections otherwise detectable only by RNA assays. The San Francisco Public Health Laboratory evaluated the performance of the Abbott Architect HIV Assay (Abbott Diagnostics) using a panel of specimens from recently infected individuals detected by RNA testing. They showed that the Architect fourth-generation assay detected 89% of infections previously found by RNA screening. The same group used this panel of recent and acute specimens to evaluate the Determine® rapid point-of-care antigen-antibody test (Alere, CO, USA). The data showed that the rapid test possesses antibody detection equal to or slightly better than third-generation laboratory antibody tests. It could detect antigen in some antibody-negative specimens, but it was not as sensitive for p24 antigen before seroconversion as the laboratory-based fourth generation tests.
New technology
The primary focus of upcoming HIV diagnostic developments appears to be near-patient testing. A presentation by the CDC described a novel rapid test for HIV p24 antigen and antibody using magnetic immunochromatographic tests. This method seemed rather powerful for point-of-care use, as it was able to detect and differentiate HIV-1 and -2 antibodies in addition to p24 antigen. While the threshold for detection of low levels of antigen appears to remain a challenge, the machine-read magnetic indicator makes the test much less subjective than visual interpretation. Exciting news from the NIH and the Henry Jackson Foundation described several efforts that are under way to develop point-of-care HIV RNA tests for both diagnosis and therapeutic monitoring. One poster presentation by Wave 80 Biosciences (CA, USA) also showed a 2-h point-of-care HIV RNA test with a threshold of detection of 10,000 copies.
FDA scientists presented data on an Europium nanoparticle-based immunoassay for HIV p24 antigen. Although not suitable for point-of-care, the assay is as much as two orders of magnitude more sensitive than existing p24 ELISAs. This might pave the way for ultra-sensitive and less expensive alternatives to RNA assays for virus detection.
Closing
The meeting ended with an open discussion of topics presented throughout the conference. A major development during this discussion was a preliminary proposal of a novel recommendation for a new laboratory testing algorithm by Bernard Branson from the Division of HIV/AIDS Prevention at the CDC. The algorithm would include screening for infection by the most sensitive available HIV-1/2 immunoassay (preferably an antigen/antibody combination test, once one is FDA-approved, although still valid for use with a third-generation assay). Specimens that are reactive on this screening test would be subjected to an antibody test that differentiates HIV-1 from HIV-2. Specimens reactive for antibodies on both tests would be considered positive for either HIV-1 or HIV-2 antibodies. Specimens negative for antibody on this second test would then be tested for HIV-1 RNA. Such an algorithm, when fronted by a fourth-generation antigen–antibody assay, could confirm HIV antibody positivity, discriminate HIV-1 from HIV-2 and detect most acute HIV-1 infections. With existing and prospective HIV-1/2 differentiation tests, turnaround time would be less than that with western blot (as illustrated by the University of Washington presentation) and the laboratory could resolve most positive specimens (for HIV-1 or HIV-2) with only two tests. The third, more costly RNA test would be invoked only for those antibody-negative specimens likely to represent acute infection (unlike RNA-pooling strategies that require testing all antibody-negative specimens). This shifts the concept of HIV-testing algorithms to one that is more sensitive than algorithms dependent on western blot, which were designed more to optimize specificity (and which miss early infections detected by today’s more sensitive screening tests).
Few meeting attendees offered any significant objections to the proposed algorithm. A concern, however, was voiced that the proposed change might result in the loss of western blots from the commercial market. Some clinicians find the blot useful for resolving diagnostic dilemmas and for monitoring aspects of the progression in immune response. Most agreed that while western blots may soon cease to be the gold standard for confirmation, their availability should continue as clinical tools for specific applications.
One notable aspect of the proposed algorithm is the dearth of discriminatory HIV-1 and HIV-2 antibody tests on the market (one such test had FDA approval at the time of writing, compared with two FDA-approved western blots). However, manufacturers are likely to seek approval for more products fitting this description.
The 250 conferees represented varied disciplines. Diagnostic manufacturers accounted for 43% of conference attendees, while members of public health departments (most directly involved with laboratory testing) made up 19%. Federal employees (CDC, FDA and Department of Defense) accounted for 14% of the attendees, and the remaining 24% were from academic institutions, hospitals and private diagnostic laboratories.
The 2010 HIV Diagnostics Conference successfully covered a broad range of topics pertinent to its attendees. The mix of government, corporate and academic scientists allowed for a meaningful exchange of data and ideas that culminated with a glimpse of upcoming products and diagnostic strategies likely to result in new federal recommendations on the horizon.
Disclosure
The findings and conclusions in this publication are those of the authors and do not necessarily represent the views of the CDC or the San Francisco Department of Public Health. Use of brand names is for identification purposes and does not imply endorsement by the US Department of Health and Human Services.
Financial & competing interests disclosure
Mark W Pandori has received honoraria in the past from Bio-Rad and Abbott Molecular. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
No writing assistance was utilized in the production of this manuscript.