The merits of measuring HIV particles in the blood as a way to predict a patient's ability to fight off the disease have been challenged.
HIV attacks immune cells
A study involving 2,800 people suggests measuring viral load is a much less reliable way to predict loss of key CD4 immune cells than previously thought.
The finding may lead to a rethink on asssessing when to start treatment.
The study, led by Case Western Reserve University, appears in the Journal of the American Medical Association.
Predicting disease progression is crucial in the treatment of HIV-positive people - in particular the decision about when to start antiretroviral therapy (HAART).
HAART has been credited with saving millions of lives, but it can cause potent side effects, and so doctors do not like to start using it until absolutely necessary.
Current treatment guidelines advise doctors that measuring viral load can be one way of assessing when to start treatment for some patients - although in the UK actual CD4 count is deemed to be the most important factor.
HIV specifically targets CD4 cells, a type of white blood cell, and as they decline in number the body loses the ability to fight off infection, raising the risk of complications associated with Aids.
The theory is that the higher the viral load, the faster CD4 cells will be lost.
However, the latest study found viral load explains only about 5% of the variation from person to person in the rate of CD4 cell loss.
This suggests CD4 depletion cannot be viewed as a simple consequence of the amount of virus circulating in the blood.
Instead, the findings suggest the factors governing disease progression are rather more complex, and may include damage that HIV is able to inflict on the immune system in an indirect way.
Lead researcher Dr Benigno Rodriguez said: "The results of this study may have profound implications in our understanding of how HIV causes disease and in our approach to the management of HIV-infected patients."
The researchers used a sophisticated statistical modeling technique to assess viral load and CD4 cell loss in more than 2,800 patients with HIV who were not receiving treatment.
Edwin Bernard, editor of AIDS Treatment Update, said it was known that CD4 decline varied enormously between people with a similar viral load, and that treatment guidelines had begun to play down the importance of the measure as a diagnostic tool.
"This study provides important new information to help us understand exactly how much viral load can predict the need to start treatment on an individual level."
He said it was clear that a variety of genetic and immunological factors affected an individual's response to HIV infection - and ways to measure these factors would become key.
However, he said the study did not question the value of viral load measurements for assessing how well antiretroviral therapy was working.
Mary Lima, of the HIV charity Terrence Higgins Trust said: "The immune system is very complex, so more research is needed to clarify how the disease progresses and when medication would be most effectively started."