May 5, 2010

START Wars

Project Inform is not the only group to struggle with the question of when to start treatment. Even the panel of experts who revised the DHHS treatment guidelines were somewhat divided—if not over their final recommendation, then certainly about how strongly they supported it. Though more than two thirds of the DHHS panelists who revised the treatment guidelines voted to change the CD4 starting point (from 350 and below to between 350 and 500), there wasn’t universal agreement on the strength of the recommendation. Fifty-five percent of the panel voted that the recommendation be “strong,” and 45 percent voted that it be “moderate.”

The rationale behind recommending treatment for those with up to 500 CD4s comes primarily from two large cohort studies. The North American AIDS Cohort Collaboration on Research and Design (NA-ACCORD) study and the ART Cohort Collaboration (ART-CC) study suggested that people who waited to start treatment until CD4s dropped below 350 faced a higher risk of premature death, from any cause, than people who start treatment at a CD4 count above 350.

Further evidence in support of early treatment—again from cohort studies—suggests that uncontrolled HIV replication might be associated with a number of illnesses not traditionally associated with AIDS. These include non-AIDS-related cancers, cardiovascular disease, liver disease, kidney disease and immune inflammation.

Simon Collins of HIV i-Base, an HIV treatment education and advocacy group in London, argues that these data need to be interpreted with a critical eye. Not only are the observed differences marginal, but the data reported thus far only come from studies with important limitations. 

“The few studies involving patients starting at 350 and 500 show both options are very safe and very effective,” Collins says. “In [NA-ACCORD], 19 out of 1,000 people died in the first year of treatment if they started at 350, compared with 16 out of 1,000 people who started above 500. The absolute risk of death in both groups was very small. The difference in the absolute risk between the groups—0.3 percent—could easily have been due to confounding factors [factors unrelated to HIV] that [cohort study limitations] can’t account for.”

As for people living with HIV with CD4s above 500, the guidelines’ panelists were again split down the middle, this time as to whether to recommend treatment or not. Whereas NA-ACCORD did suggest a benefit associated with very early treatment, ART-CC did not. Even with NA-ACCORD, the guidelines’ authors themselves note important limitations, “including the small number of deaths”—which makes detailed comparisons extremely difficult—“and the potential for unmeasured confounders that might have included outcomes independent of antiretroviral therapy.”

“This brings us to why randomization is key to the quality of the evidence,” Collins says, pointing out that cohort studies lack this tried-and-true scientific device that can help neutralize confounding factors in studies. “Randomized trials generally get a balance of all factors in the two groups being compared. This includes the things we think might affect the results, but also the things we have no control over, like genetics or personal behavior.”

Pages: 1 | 2 | 3 | 4 | 5

Search: START, randomized studies, HIV treatment, antiretroviral therapy, Project Inform, Bob Huff, Mark Milano, Simon Collins, controversy

Scroll down to comment on this story.