Treatment interruption continues to be controversial strategy in HIV care

CROI studies show mixed results

Despite national media headlines announcing the end of treatment interruption as an antiretroviral strategy, the most recent studies show mixed results, and so the research probably will continue.

"The problem with the drug interruption studies is you can't generalize them because the questions asked in the trials were very different," notes Kathleen Squires, MD, professor of medicine and director of the division of infectious diseases at Jefferson Medical College in Philadelphia, PA.

For instance, the widely-publicized results from the SMART study, which involved 5,472 patients in 33 countries, showed that CD4-guided episodic antiretroviral therapy use is associated with increased risk of HIV disease progression and death when compared with continuous antiretroviral treatment.1

Likewise, a randomized structured treatment interruption trial in Abidjan, Cote d'Ivoire, found that CD4-guided therapy (CGT) with introduction/interruption thresholds of 250/350 CD4 cell counts resulted in twice the serious morbidity rate as continuous therapy, mainly because of invasive bacterial diseases.2

After a data safety monitoring board (DSMB) reviewed data in an interim analysis in October, 2005, the DSMB recommended stopping the CGT arm prematurely, and so the arm was discontinued.2

An HIV Outpatient Study (HOPS) that addressed early, uninterrupted antiretroviral treatment suggests that the use of HIV antiretroviral therapy without interruptions may reduce the incidence of the complications of renal insufficiency, peripheral neuropathy, and lipoatrophy.3

But other studies presented at CROI had different conclusions.

An ACTG 5170 study concluded that antiretroviral therapy interruption was generally safe among its cohort of HIV-infected patients, who had been on antiretroviral therapy for at least six months and had viral loads fewer than 55,000, as well as CD4 cell counts over 350, both prior to starting therapy and at the study entry.4

"To summarize, we found a single treatment interruption resulted in low progression rates at two years," says Daniel J. Skiest, MD, director of adult HIV services at Baystate Medical Associates in Springfield, MA.

While patients didn't do well with treatment interruption on the SMART study, there were several studies where patients did fine with treatment interruption, Skiest says.

"It depends on who you choose as subjects and how you choose to stop therapy," Skiest says. "You can't choose people who had AIDS before or a low CD4 cell nadir."

In Skiest's study, the patients did best who had relatively high CD4 cell counts and who had no AIDS complications, and they have to be watched carefully, he says.

"There is not a straight black and white answer because the studies had different conclusions," Skiest says. "But all of the studies had different entry criteria and different interruption strategies."

The Staccato trial found few resistance mutations during scheduled treatment interruptions. The patients randomized to be on the scheduled treatment interruption arm were taken off treatment as long as the CD4 count exceeded 350.5

A pediatric treatment interruption study showed that a subset of pediatric patients who were treated with increasing periods off therapy had improved virologic control despite the periods off therapy. The 14 children were studied on a repeated supervised treatment interruption, with periods of interruption that began with three days and increased by two days for each cycle, resulting in a treatment interruption that eventually exceeded 27 days in length and included 13 or more interruptions.6

The ACTG 5170 study was begun because investigators were interested in outcomes of patients who had started antiretroviral therapy under the previously more aggressive treatment guidelines, Skiest says.

"Several years ago, they started therapy when a patient's CD4 cells were below 500, and now it's when the CD4 cells are 200 to 350," Skiest says. "The question was 'Is it safe to stop therapy in these patients?'"

Such long-term patients sometimes have unpleasant side effects from their antiretroviral treatment, and they have treatment fatigue, Skiest notes.

Investigators wanted to see what would happen when therapy was stopped and whether this would have an impact on the patients' quality of life.

"We also wanted to see if we could come up with some predictors that we could measure before they stop therapy that would predict how they would do after stopping therapy," Skiest says.

The prospective, single-arm, observational study looked at 167 patients at 31 sites in the United States. Patients had to be on continuous antiretroviral therapy for at least six months, and their viral load had to be under 55,000, he says.

"We purposely chose people who were doing well and may not need therapy because they had such good numbers," Skiest adds.

"So when they entered the study, they stopped all therapy, and we followed them for 96 weeks," Skiest explains. "They could restart at any time if they wanted to or if their physician wanted us to start therapy again."

If a patient's therapy was restarted, investigators followed the patient's outcomes for 24 weeks, he adds.

At the study's entry, the median CD4 cell count was 833, and 71 percent of patients had viral loads of less than 50 copies with a median nadir in CD4 count of 436, Skiest says.

"The median viral load pre-therapy was 22,611, and that's not very high," Skiest says. "That's reflecting a population of patients who may not need therapy."

The CD4 count nadir was predictive of response to treatment interruption, Skiest says.

"If you had a nadir less than 400 cells than those patients reached one of the end points faster and were almost twice as likely to reach that end point," Skiest says. "We also found that the viral load at the time of the study entry, if it was over 400, then that predicted a faster time to reaching the same end point."

Only 18 percent of the patients enrolled in the study had viral loads over 400 copies, he notes.

In all, 46 patients restarted therapy, and five died, although investigators couldn't draw any conclusions from the deaths, Skiest says.

"Most who died had high CD4 counts when they died, and four had low viral loads," he says. "Three had a significant history of coronary artery disease, and we thought two or three of them had died from a cardiac event, but that was a best guess because there was only one autopsy."

Researchers will continue to analyze the results, including the quality of life and neurocognitive measures, Skiest notes.

Research into treatment interruption likely will continue because there are a variety of possible advantages to the strategy, and not all of them have been adequately explored, Skiest says.

Some of the theories have been shown to be incorrect, including the hypothesis that treatment interruption will allow wild type virus to re-emerge in people who have resistant virus, Skiest says.

"That hasn't worked," he says. "Another theory is that treatment interruption will boost the immune system, and that doesn't work."

It's still open to debate whether a public health system could spare therapy in patients who are doing well and save considerable money, Skiest says.

"Some studies say you can save 50 percent of antiretrovirals, which is a lot of money for a developing world," Skiest says. "So if you can do it safely and carefully choose patients based on our study and other studies at the CROI meeting, then I think that would be a reasonable strategy."

Squires says this potential benefit of treatment interruption has a major drawback: "The problem with that is many patients in resource poor settings tend to have advanced HIV disease by the time they present with HIV infection," Squires says.

"My interpretation is it wouldn't be a good idea to cycle people on antiretrovirals," she says. "But you have to have a balance between the availability of the drug, cost of drugs, and the ability to get drugs to the people who need them."

In the United States, there have been a lot of patients who started HIV therapy under the more aggressive guidelines and then they stopped therapy and are doing okay, Skiest says.

"There aren't as many patients now who would stop therapy because our guidelines are much less aggressive now," Skiest adds. "We don't start until the CD4 count is lower."

So if treatment interruption is used here it likely will be because the patient requests it for relief from toxicity of drugs and treatment fatigue, he says.

"They might take a short break," Skiest says. "But you have to be careful about stopping therapy, depending on what regimen the patient has because there's some concern about resistance developing, like with non-nucleoside reverse transcriptase inhibitors (NNRTIs), which have a longer half life than other drugs."


  1. El-Sadr W, Neaton J. Episodic CD4-guided use of ART is inferior to continuous therapy: results of the SMART study. Presented at the 13th Conference on Retroviruses and Opportunistic Infections, held Feb. 5-8, 2006, in Denver, CO. Abstract: 106LB.
  2. Danel C, et al. The CD4-guided strategy arm stopped in a randomized structured treatment interruption trial in West-African Adults: ANRS 1269 Trivacan trial. Presented at the 13th Conference on Retroviruses and Opportunistic Infections, held Feb. 5-8, 2006, in Denver, CO. Abstract: 105LB.
  3. Lichtenstein K, et al. Early, uninterrupted ART is associated with improved outcomes and fewer toxicities in the HIV Outpatient Study (HOPS). Presented at the 13th Conference on Retroviruses and Opportunistic Infections, held Feb. 5-8, 2006, in Denver, CO. Abstract: 769.
  4. Predictors of HIV disease progression in patients who stop ART with CD4 cell counts >350 cells/mm3. Presented at the 13th Conference on Retroviruses and Opportunistic Infections, held Feb. 5-8, 2006, in Denver, CO. Abstract:101.
  5. Ananworanich J, et al. CD4-guided scheduled treatment interruptions: low incidence of resistance mutations in the staccato trial. Presented at the 13th Conference on Retroviruses and Opportunistic Infections, held Feb. 5-8, 2006, in Denver, CO. Abstract: 622.
  6. Borkowsky W, et al. Repeated supervised treatment interruption with progressive increases in "off-treatment" duration results in enhanced virologic control in a subset of pediatric patients. Presented at the 13th Conference on Retroviruses and Opportunistic Infections, held Feb. 5-8, 2006, in Denver, CO. Abstract: 19.