There has been a lot of interest in what some people call "drug holidays" -- that is, stopping anti-HIV medications for limited periods of time. Since there is no clear answer as to whether this is beneficial or even safe, Richard Jefferys of Treatment Action Group presented an overview of some of the most recent important studies.
Structured Treatment Interruptions (STIs) fall into three main categories:
A study at Johns Hopkins University looked at 101 people who stopped HAART (Highly Active AntiRetroviral Therapy). The most common reasons were having started treatment too early according to the revised treatment guidelines, side effects, or problems with adherence to therapy.
Two-thirds of those who stopped are still off HAART after an average of 74 weeks, and the average CD4 count is 508, with a viral load of 43,000. One-third (33 people) restarted therapy due to rising viral loads, declining CD4 counts or because they were nervous about being off therapy. All but two of these individuals have now suppressed their viral load to less than 400 copies on HAART. The best predictor of having to restart was a low CD4 count prior to starting HAART, although about 40% of people who had CD4 counts below 350 prior to starting HAART remain off therapy.
The study calculated that people who started HAART with CD4 counts over 350 and viral loads below 55,000 could stay off therapy for 5.7 years (or possibly longer) before reaching a CD4 count that was below 200. So, "If you didn't need HAART then, you may not need it now." Since this study did not randomize people to stop or continue treatment, it couldn't prove anything, so other trials, such as the CPCRA's SMART study, are being conducted to look into this more carefully.
A trial done in Argentina randomized people on HAART who had never had a CD4 count below 350 or a viral load above 60,000 to either stop or continue HAART. The early results show that after six months off treatment, CD4 counts are still greater than they were before HAART was started. Some people had a big jump in viral load when they stopped therapy, followed by a decline; others had a smaller increase and then a decline. But everyone's viral loads stabilized about twelve weeks after stopping treatment and stayed within one log of their viral load before HAART was started. Surprisingly, there was no significant change in lipid levels, although there was a significant increase in hemoglobin, and a significant decrease in lactic acid levels.
Concerning the effect of STIs on drug side effects, results look good when it comes to lipid levels. One study looked at 25 people who stopped HAART when they had viral loads below 50 and an average CD4 count of about 400. So far, the average time off HAART is 11 months, and only four people out of the 25 have had to restart treatment. The study has seen significant decreases in cholesterol and triglycerides, but increases in insulin levels. Another study found significant decreases in cholesterol and triglycerides in 26 men on HAART who underwent a brief, one to two month treatment interruption. This study saw no change in glucose or insulin, however.
Finally, a look-back study of HIV-related illnesses during the first three months off therapy in 247 people found only 18 clinical events, none of which were life-threatening -- eleven cases of thrush, two herpes zoster infections, one herpes simplex outbreak, one hairy leukoplakia, one case of salmonella, and two cases of bacterial pneumonia. The average CD4 count was lower in people with clinical events compared to those without any (261 vs. 499).
Another approach being studied in people who had controlled their virus while on HAART is a "7 days on/7 days off" cycle. So far, in a pilot study of only ten people, no one has a detectable viral load, even though some people have repeated this cycle for up to three years. This is a very small study, however, and larger trials need to be conducted to verify whether this is safe.
The French GIGAHAART study looked at 62 people who had taken all classes of anti-HIV drugs for a long time (six to seven years) and who had multiple resistance mutations. They had advanced HIV disease, with an average CD4 count of 26 and an average viral load over 100,000. Study participants were randomized to either begin GIGAHAART immediately or after an 8-week break. GIGAHAART consisted of taking six to nine antiretrovirals.
In this study, those who stopped treatment before restarting saw a clear benefit. After 24 weeks (six months), they had an average viral load drop of 1 log compared to a 0.29 log drop in those who began treatment immediately, and their average CD4 count increase was 51 compared to 8 in the group who began GIGAHAART immediately. There was one case of tuberculosis and one of recurrent herpes zoster in the treatment interruption group compared to one fatal case of lymphoma, one diagnosis of PCP and one dual diagnosis of cryptosporidiosis and microsporidiosis in the group that went immediately into GIGAHAART. This study showed a surprising benefit to an STI, so follow-up will continue to see whether or not the benefit is real.
Steven Deeks, M.D., an expert on STIs, offered his personal opinions at the International STI Workshop in March of 2002:
And, of course, there is always the option of trying to access to experimental drugs through clinical trials or expanded access programs.
In a study by Bruce Walker, fourteen people who had recently become infected with HIV (acute infection) took HAART before they developed antibodies to HIV. They then did between one and three STIs. Eight of the fourteen have maintained viral loads below 5,000 up to three years after infection. Only one of the fourteen is still on continuous HAART. Those who were able to control their viral loads while off therapy had HIV-specific CD4 and CD8 responses. But it's hard to interpret these data because everyone did HAART followed by STIs. It's possible that they might have been just as successful if they had never taken HAART to begin with.
The Swiss-Spanish Interruption Trial (SSITT) studied 133 people who were taking HAART and had viral loads below 50. They interrupted treatment four times (two weeks off, eight weeks on), and then stopped treatment after 40 weeks. Sixty-six of the 133 people withdrew from the study when their viral loads did not go below 50 once they restarted treatment or because their viral loads went too high after they stopped.
After a year, 23 of those remaining had viral loads below 5,000 -- nineteen of those remained off treatment after 20 months, and 41 people total were still off treatment at that time. The good news is that no one developed an opportunistic infection, and only one person developed drug resistance. The bad news is that there was little evidence of auto-immunization, and there was one case of HIV being transmitted to an HIV-negative partner during the treatment interruption.
More studies are continuing with differing schedules and with therapeutic immunizations to increase the chance of an immune response while off treatment.
Several studies of repeated STIs have found cases of drug resistance -- the risk is less likely in the case of a single extended interruption. Another concern is that "archived" drug mutations can exist even in people who have viral loads below 50. An STI could cause the mutations to re-appear. So in the end, the risks of STIs remain unclear.