Treatment Interruptions: SMART or Stupid?

Spring 2006

Ten years ago, when HAART (highly active antiretroviral therapy) made its debut, the theory behind using a combination of HIV drugs seemed easy enough to understand. Preliminary data suggested that HIV could be cured -- "eradicated," as experts were prone to say -- if people on HAART could simply keep their viral loads undetectable for a little more than two years. Two years of treatment? No problem.

Unfortunately, within a few years several studies confirmed that even the most powerful combinations of the available HIV drugs were unable to completely stop HIV from reproducing in the body. As a result, the hope of eradicating HIV gave way to a much more sobering reality: once someone needs HIV drug treatment, that person will likely need to stay on therapy for the rest of his or her life.

Ten years later, real-world experience has shown how difficult it can be for some to stay on HAART for very long periods of time. HIV meds can cause a number of long-term side effects, including liver problems, body-shape changes, and increased lipids (such as cholesterol and triglycerides). Because of this, researchers soon began looking into a treatment technique called "structured treatment interruptions," or STIs for short.

STIs have been studied in a number of different ways. They have been tried in people who achieve high CD4 counts after many months or years on treatment; in people experiencing long-term side effects of HIV meds; in people with HIV drug resistance (with the hope that stopping therapy might help their virus switch to a strain that is sensitive to available drugs); and to boost the immune system (using STIs to boost the activity of the HIV-specific CD4 cells that can help control HIV). Results from these studies, however, have been mixed.

Another approach that has been explored is the possibility of treating HIV like other chronic diseases: starting therapy when the immune system shows signs of damage or when a patient experiences symptoms of HIV disease, and stopping it when their health improves. This STI approach was explored in the SMART (Strategies for Management of Antiretroviral Therapy) study, a large (6,000 people) and long (almost 9 years) clinical trial.

Not expecting results for several years, many healthcare providers, advocates, and people living with HIV were taken by surprise when it was announced in January that SMART was being stopped after only 14 months, due to a high rate of disease progression in those who were assigned to STIs. But do the SMART results signal a death knell for this approach and other types of STIs? Let's take a closer look.

SMART enrolled people with CD4 counts above 350, who were either on treatment or who had not yet started therapy. They were randomized to one of two groups: a continuous treatment (CT) group that did not stop HIV treatment, or an STI group that stopped treatment until the CD4 count fell below 250, restarted when the CD4 count rose above 350, stopped again until it fell below 250, and so on.

The researchers conducting SMART hoped that repeated STIs would allow people to maintain CD4 counts that would be high enough to keep them healthy while at the same time reducing the amount of time on treatment and lowering the risk of side effects.

On January 10th, when the study's DSMB (Data Safety and Monitoring Board) recommended halting SMART, the study had enrolled 5,472 patients. While the study is technically still being conducted, no new patients are being enrolled and patients in the STI group -- especially those who had tried and failed other anti-HIV drugs in the past -- have been advised to restart treatment and to remain on therapy.

Dr. Wafaa El-Sadr, Chief of the Infectious Disease Division at Harlem Hospital in New York and a leading SMART researcher, presented the data at the recent Retrovirus conference in Denver. She reported that there were 164 "events" during the 14 weeks of study follow up. These events were defined as either death or the development of a serious AIDS-related condition or another serious complication (such as a potential side effect).

There were 47 events in the CT group, and 117 events in the STI group. In other words, people on STIs had 2 1/2 times the risk of a serious event. 0.6% of patients in the CT group progressed to AIDS, compared to 0.9% of patients in the STI group. (Esophageal candidiasis was the most common AIDS-related sign of disease progression among people on STIs.) And 0.9% of patients in the CT group died, compared to 1.7% of people in the STI group.

Cause of DeathSTICT

The causes of death in the study were confusing. Only a small number of deaths were due to AIDS-defining conditions. And while there were slightly more deaths in the STI group from some causes (as seen in the table above), the real difference came in the "Other" and "Unknown" categories. "Other" refers to deaths stemming from violence or accidents. But until we know what caused the deaths in the "Unknown" category, it will be hard to draw final conclusions from the study.

Another surprise: the SMART researchers were hoping that STIs would lead to fewer complications that can be caused by HIV treatment, such as heart attacks, stroke, coronary artery disease, kidney problems, and liver damage. But they actually found more complications in the STI group. According to Dr. El-Sadr's report, 2.1% of patients on STIs experienced a serious complication, compared to 1.4% of patients in the CT group.

Many STI studies have reported that the lower a person's CD4 "nadir" (the lowest their CD4 count has ever been), the more likely it is that someone will progress while attempting an STI. This was not the case in SMART -- people with both low and high CD4 nadirs experienced "events" during the study. What's more, patients who had viral loads below 400 at the time they stopped therapy were more likely to have an event than those who stopped therapy with higher viral loads. (Perhaps the inflammation associated with viral load rebounds after a long period of suppression may be harmful, but this is only conjecture.)

So SMART's conclusion, based on the data presented so far, is clear but limited: STIs using the approach of stopping treatment when CD4 counts are higher than 350 and restarting when it drops below 250 are inferior to continuous treatment -- the current standard-of-care.

But questions were raised at the conference. A sizeable number of people had a history of an AIDS diagnosis (24%), and many (95%) had tried and failed other treatments in the past. Would a study involving people who had a healthier history and less resistance to anti-HIV treatment yield different results? Were the stop and start points (250 and 350) too close together? Could the time since the CD4 nadir have an effect on the safety of an STI?

Making sense of the SMART data will be tricky. Similar trials presented at the conference showed conflicting results. For example, in the STACCATO trial, a similar study reported at the conference, people doing STIs who restarted treatment at a higher CD4 count-350, compared to SMART's 250 -- were no more likely to die or experience an AIDS-related complication, compared to those who remained on continuous therapy. But another study reported at the conference (the TRIVACAN study), mirrored the results of SMART. It, too, used a CD4 count of 250 to restart patients on treatment.

While SMART, TRIVACAN, and STACATTO all used patients' CD4 counts to determine when treatment should be stopped a restarted, other STI approaches have been tried, including those using weekly and monthly timeframes instead. For example, the Italian ISS PART study compared 137 people on continuous therapy to 136 people doing increasingly long STIs (of one, two and three months, each followed by three months on treatment), and had positive results: while the STI group had a slight drop in CD4 counts after two years (a decrease of 28 cells compared to no decrease in the CT group), 91% of the STI group and 92% of the CT group had viral loads below 400.

The French WINDOWS study compared CT to a two-months-on, two-months-off STI approach in 403 people. After 21 months, there were no AIDS-defining events in either group. While more people in the CT group had CD4 counts above 450 (92% compared to 75% of people doing repeated STIs), the development of resistance was similar in both arms (17 in the STI group and 14 in the CT group).

So perhaps the only real lesson from the SMART study is that using CD4 counts of 350 and 250 to repeatedly stop and restart therapy may not be such a good idea, especially in people with a history of advanced disease. The STACCATO study suggests that episodic treatment, using CD4 counts as a guide, may be possible if a higher CD4 count is used to restart treatment. And the ISS PART and WINDOWS studies suggest that structured weekly or monthly timetables for stopping and restarting treatment may also be worthwhile.

In the end, as often happens with studies that are designed to give a "final answer" to a major question, the conclusions of these studies yield even more questions than answers. For the near future, the message for people with HIV who want to take a break from their meds remains the same: STIs are risky at best, and should not be attempted without a careful evaluation of the risk and benefits, and certainly not without the guidance of an HIV expert or within a clinical trial.

Tim Horn is Senior Writer and Editor of Mark Milano is a longtime HIV treatment activist and Editor of ACRIA Update.

This article was provided by AIDS Community Research Initiative of America. It is a part of the publication ACRIA Update. Visit ACRIA's website to find out more about their activities, publications and services.


The content on this page is free of advertiser influence and was produced by our editorial team. See our content and advertising policies.