Alternative to viral load flops in West African field trial

It’s been noted that the impact of highly active antiretroviral therapy (HAART) in 1995 might not have been so dramatic had it not been accompanied by a simultaneous breakthrough in assay technology called HIV-1 RNA PCR, which allowed the drop in viral load caused by the new drugs to be rapidly detected. The HIV RNA viral load test is now used routinely to monitor the initial response the HAART and to give an early warning when viral suppression starts to falter so that regimen changes can be made before drug-resistant mutations get out of control. The other main assay in the clinician’s toolkit, the CD4 test, tracks a person’s immune status, telling when it’s time to start therapy and reporting on the slow process of immune recovery after successful viral suppression has been achieved. But both of these tests involve sending blood to qualified laboratories where trained technicians use sophisticated equipment to obtain reliable and reproducible results. The cost of these tests can range from $60 to well over $100 each.

In the developing world, where the burden of HIV infection is greatest and resources for fighting the epidemic scarce, recent dramatic reductions in the cost of antiretroviral drugs are promising to make HAART available to millions of people who would otherwise die without it. Yet while drug prices have plummeted and access is improving, the high cost and unavailability of diagnostic assays to monitor therapy threatens to limit the impact of HAART in sub-Saharan Africa and elsewhere.

In particular, some experts worry that with no way to monitor the success of therapy there is a danger that an epidemic of drug resistance could erase the gains made by introducing drugs without affordable diagnostics. Also, without the ability to detect treatment failure early, it may not be possible to deliver the sustained improvement in survival and health that has occurred since 1996 in the richer nations.

There are a number of ongoing efforts to simplify the measurement of HIV RNA and CD4 counts by using alternate technology that doesn’t require expensive lab equipment or by relaxing the precision of the tests so that they simply return a pass/fail result instead of a numerical readout, but these advances are years away.

Another avenue of research into affordable diagnostics has been to develop alternative markers of treatment success that don’t rely on HIV RNA. One of these methods involves looking at blood levels of an HIV protein called p24, which is produced in abundance when HIV replicates. The detection of specific proteins in the blood is a simpler process than detecting RNA, and if p24 proteins levels were found to correlate with the reduction of viral load after treatment or to predict viral resurgence after treatment failure, then the cost of monitoring HAART could be brought down to about $19 per test. Although p24 monitoring was used in the pre-HAART days as an experimental measure of viral activity, it was never found to be particularly reliable. Yet, advances in the underlying technology and the market demand for more affordable lab tests in the developing world have put p24 back in focus as a potential diagnostic. Recently, a new generation enzyme-linked immunosorbent assay (ELISA) for p24 has shown good correlation with HIV RNA viral load in its ability to quantify HIV under highly controlled conditions using common Western strains of HIV. But the value of the technology for use in Africa with non-subtype B strains of HIV has been an open question.

A study by Dominique Bonard and colleagues from the Center for AIDS Diagnostics Research in Abidjan, Cote d’Ivoire published in the November 1^st issue of the Journal of the Acquired Immunodeficiency Syndrome reports on a head-to-head comparison of p24 ELISA with HIV RNA in treated and untreated Africans with HIV. Unfortunately, the promise of p24 as a cheaper, but no less reliable, marker of treatment response than RNA viral load is not supported by the data they present.

To evaluate the fundamental sensitivity and specificity of the competing assays, 192 plasma samples from HIV-positive (117) and HIV-negative (75) persons were tested using each method. While both the RNA and p24 tests each correctly identified about 96 percent of the HIV-positive samples, the p24 test mistakenly returned borderline false-positive results for 4 of the 75 HIV-negative samples. While the sensitivity of the test (the ability to detect small amounts of p24) was in line with previously published results, the specificity (the ability to detect p24 and nothing but p24) was considered suboptimal. An analysis of the tests’ ability to quantify HIV only found a significant correlation at viral loads in the neighborhood of 100,000 copies, far above the useful range for detecting treatment response or failure.

Next, the investigators evaluated the ability of p24 to report response to therapy in persons starting HIV treatment for the first time in an open label study of indinavir and Combivir. Fourteen patients were selected, including the first seven to have a successful treatment response and seven others who experienced viral rebound during the trial. While HIV RNA tests were performed at baseline and at scheduled visits during the study, the p24 test was performed at a later date on frozen samples collected at those visits.

The results were striking in the lack of information provided by p24 compared to HIV RNA. While each of the seven patients with a successful response to HAART had an evident drop in HIV RNA by month 2, p24 was either unchanged or only weakly reflective of the impact of therapy. For the seven patients with viral rebound as detected by HIV RNA, p24 was mostly unresponsive; giving no indication that viral replication had become more active. Finally, in every case where HIV RNA had become undetectable (less than 200 copies/mL) the p24 test was still reporting measurable quantities of the protein.

The failure of p24 to respond to changes in viral load might possibly be related to a diminished ability of the test to recognize non-B clade HIV. Alternatively, the retrospective evaluation of frozen samples for p24 may have introduced some unknown factor that affected the results. More likely, though, is that the presence of p24 simply does not mean the same thing that finding RNA does. For example, some viral proteins may continue to be produced despite suppression of viable HIV. While there may yet be a role for p24 in detecting a productive HIV infection in newborns, this study strongly suggests that measuring p24 offers little guidance for monitoring response to antiretroviral therapy.

Further information on this website

Monitoring where resources are limited - overview of HIV diagnostics in resource-limited settings.