"Some clinicians have wondered whether in these cases they should change the patient's drug regimen. This could be a problem, because usually the patient has been on what has been selected as the optimal drug combination for his particular situation. Any change will lead to use of a combination that's less optimal, so you don't want to change unless you need to," he said.

One of the significant problems in detecting drug resistance is that genotyping tests are only sensitive down to levels of about 1,000 copies per milliliter. So, without the ability to test for resistance at 120 copies, clinicians are left in the dark about the significance of the blips with regard to resistance, Siliciano said.

Siliciano and his colleagues hypothesized that the small increases in viral load represented by the blips were only random statistical fluctuations in measurement of a viral load that was clinically well managed. Such fluctuations would have no clinical significance, they theorized.

They also believed that they would not detect any evidence of developing drug resistance based on their considerations of the likely source of the blips. "We had shown that there is a latent reservoir for HIV in resting CD4 immune cells that is not affected by drugs," he said. "This stable reservoir will continue to spit out low levels of virus. This is a likely source of low levels of virus in the blood, rather than ongoing viral replication, which might lead to drug resistance."

To confirm their hypotheses, they tested the viral levels in 10 patients every 2-3 days -- a much more frequent interval than the standard practice of testing once every several months. To determine whether the blips were statistical fluctuations, they submitted the samples to two independent laboratories for testing.

They a
'"/>

Contact: Jim Keeley
keeleyj@hhmi.org
301-215-8858
Howard Hughes Medical Institute
15-Feb-2005