Viral load testing: which test for what? Dr Raul Scott Pereira

Dr Raul Scott Pereira MA MB BChir PhD FRCPath Senior Lecturer and Consultant in Immunology, Chelsea & Westminster Hospital, London

HIV-1 viral load in plasma is a new way of monitoring the progress of HIV disease. There are currently three commercial assays available, each of which uses a different amplification method to detect and quantitate viral RNA and measure the number of RNA copies per ml of plasma (they do not measure virus in circulating cells or total body viral load). The detection limit is continually improving and levels of 10 RNA copies may soon be both detectable and reliably quantitated.

In early disease, viral load may be detectable at several million copies per ml, but it then settles to a stable `set point', which has been found to be a better prognostic indicator than the CD4 lymphocyte subset count.

In early drug trials, a fall in viral load of 0.5 log 10 was considered evidence of therapeutic effectiveness. Recently, the aim with highly active antiretroviral therapy (HAART) has been to achieve levels below the limit of assay detection. Within these limits, the tests have become an essential adjunct to current treatment, in which they justify their expense by permitting fine control of expensive antiretroviral drugs.

Evaluation of tests

Following several months' experience with all three assays, we recently evaluated them, in collaboration with the PHLS Central Public Health Laboratory, for sensitivity , reproducibility , staff and laboratory resource requirements and test cost . The study is in preparation for publication as a Medical Devices Agency Kit Evaluation report. The assays evaluated were:


We took 120 unselected samples from adult patients attending the local clinic (Kobler Clinic, St Stephen's Centre) and analysed them by all three methods, with between- and within- assay replicates. The results, in brief, show that all three assays have acceptable reproducibility (between- and within-assay variation).

There was a detectable between-assay bias suggesting that the RT-PCR assay gave higher results than the others.

About a third of samples gave results below the level of detection with one or more assays and about 30 with all three. These have since been further tested by the ultrasensitive modifications of the RT-PCR and NASBA methods and seven were still undetectable.

Clinicians were asked to look at the results generated by the multiple assays and indicate whether any would have changed their management. Approximately 9% of results were clinically significantly different by this measure.

The hands-on time taken by each assay, in a busy laboratory with two staff testing 80 samples a week, suggests that NASBA is the most labour-intensive .

Cost per test , on a similar basis: NASBA £66, RT-PCR £63, bDNA £55.


On balance, these assays are all scientifically satisfactory; their clinical usefulness depends, however, on sensible and parsimonious use. BHIVA guidelines 1 suggest testing untreated patients 6-12 monthly and those undergoing treatment 3-6 monthly. As ultrasensitive methods become more widely available, it would seem sensible, as their range is small and cost slightly greater, to restrict their use to testing after a first negative result in a more conventional assay.

The future may lie with additional in vitro tests of viral resistance to therapy, which will allow the rapid selection of new drugs when viral load rises again. Current costs of these tests are high, but rapid, effective and rational changes to expensive drugs should make them cost-effective and, on currently available evidence, clinical response to therapy gives rise to justifiable optimism.


  1. BHIVA Guidelines Co-ordinating Committee. British HIV Association guidelines for anti-retroviral treatment of HIV-seropositive individuals. Lancet 1997; 349: 1086-1092.
(Much of the most useful current information is unpublished or in meeting abstract form and does not therefore constitute an evidence base.)