As a person living with HIV for almost 30 years, I've had a love/hate relationship with the HIV treatment guidelines put out by the Department of Health and Human Services since 1998. I loved the fact that a panel of experts, including clinicians, researchers, and people with HIV, was meeting regularly to debate and discuss HIV treatment, especially the difficult question of when to start HIV meds (since I wasn't taking them). But I hated that these "guidelines," which often were just best guesses, were treated as gospel by many doctors, including mine.
The guidelines came into being after the introduction of triple combination HIV therapy in 1996. With so many drugs to choose from, and so little data on who should take them, people were hungry for some expert guidance. I recall being told by many doctors and researchers, "You were so smart to avoid sequential monotherapy (jumping from one drug to the next), but now you really have to start combination therapy." The idea was to "hit hard, hit early" -- fight HIV with a powerful combination of drugs as early as possible in the course of the disease.
This made sense in theory (if you have a life-threatening infection, why not fight it as early as possible?), but the reality of HIV drugs back then was not so simple. Nasty side effects (one friend called full-dose Norvir the worst poison he had ever taken), ridiculous dosing schedules (Crixivan had to be taken every eight hours on an empty stomach), and the risk of resistance if even a few doses were missed. Add to this the fact that the only hard data on the drugs' ability to extend life came from studies of people with CD4 counts below 200, and you can see why people with higher counts and no HIV symptoms had a hard time with this approach.
The guidelines panel seemed to agree -- after recommending treatment for anyone with a CD4 count below 500 or a viral load above 20,000, they backtracked in 2001, advising that "many experts would defer therapy" in people above 350 with a viral load below 55,000. In 2004, the viral load threshold was raised -- even if it was above 100,000, the guidelines stated: "most clinicians recommend deferring therapy in those above 350, but some ... will treat." In 2007, viral load was removed from the equation and the guidelines became clearer: "therapy should be initiated" in anyone with a CD4 count below 350. But for people above 350, there was still doubt: "the optimal time to initiate therapy ... is not well defined."
Raising The Bar
Which brings us to last December's big change. When the panel met, they knew that their earlier decision to recommend treatment for anyone below 350 was now supported by the first major clinical study of when to start -- the CIPRA HT 001 study, done in Haiti. People with CD4 counts between 300 and 250 either started HIV treatment immediately or waited until they dropped below 200. The study was stopped early because 23 people who delayed treatment died, compared with only six people who started immediately. Since the study was randomized (people were assigned to start or wait by chance) this was the first hard data proving that 350 was no longer a guess -- it was a real benchmark of when to start.
Now the question was: should the starting point be raised even higher? Should we go back to recommending treatment to anyone below 500? It was déjà vu all over again, but with some important differences. First, the drugs today are not like the drugs in the mid-90s. Fewer side effects and easier dosing (including a number of once-a-day regimens) have addressed many of the arguments against early treatment (though of course we don't know the long-term side effects of newer drugs). Second, we know more about what HIV actually does to the body, even at higher CD4 counts. As discussed in the cover story of this issue, we're learning much more about HIV inflammation. Turns out HIV not only suppresses the immune system, it also activates it, which can lead to heart, kidney, and liver disease, and cancer. In addition, the SMART study found that people who interrupted their HIV treatment had a higher risk of death, mainly due to non-AIDS-related causes like heart disease and cancer.
Finally, cohort studies like ACCORD found that people who began treatment at higher CD4 counts -- even above 500 -- lived longer than those who waited. But cohort studies have one major flaw: people choose what to do themselves, rather than being assigned by chance, like the CIPRA study. When people choose when to start, it's hard to tell if the final result is due only to the treatment or to something else. And even though ACCORD saw a benefit to starting earlier, the difference was small -- meaning the result could have been due to other factors.
So we need a study that randomly assigns people to start or wait -- and that study has just begun. The START trial will enroll 4,000 people who have CD4 counts above 500. Half will start as soon as they drop below 500, and half will wait until they reach 350. This should give us a definitive answer on when to start -- but it won't report results until 2016.
The panel clearly thought that was too long to wait, and decided to take on the issue last December. The final vote was divided on the strength of the recommendation: 21 voted to strongly recommend treatment to anyone below 500, while 17 voted for only a moderate recommendation. Usually, the panel requires recommendations to be passed by a 2/3 majority, but in this case the recommendation was published as a split vote: 55% to 45%.
When it came to people above 500, the panel was split down the middle: 19 voted to recommend treatment, in the hope that it would lower the inflammation and non-AIDS conditions apparently caused by HIV, and 19 voted against, due to concerns about long-term side effects, the difficulty of adherence for people with no symptoms, and the risk of resistance. One panel member, James Neaton, felt that giving HIV drugs to people at low risk of disease could outweigh the "modest predicted benefit." He was quoted in the New York Times as saying, "That is why we do randomized trials."