Studies finds risks in guided treatment interruptions even at higher CD4 cell threshold

A study published in the January 11^th edition of AIDS found that “guided treatment interruptions” in people with CD4 cell counts greater than 350 cells/mm³ do not in fact reduce the overall burden of symptoms, compared to continuous antiretroviral therapy. However, patients who interrupted therapy reported better psychosocial (though not physical) quality of life. An editorial review in the same issue suggests that patients may still elect to use such strategies in certain limited circumstances, despite predominant evidence against them.

Many studies have investigated planned and supervised breaks from ongoing antiretroviral treatment (structured treatment interruptions). The intended goals are to reduce toxicities, financial cost, and the quality-of-life issues caused by ongoing, lifelong pill burden. Such structured interruptions have used either fixed time strategies (e.g. eight weeks on, eight weeks off), or have been guided by clinical markers (i.e. using fixed CD4 cell count levels to trigger on- and off-therapy intervals). These latter have been termed guided treatment interruptions, or GTIs.

Unfortunately, the large-scale (5,000 patient) SMART (Strategy for Management of Antiretroviral Therapy) study of CD4 count-guided treatment interruptions showed a significant increase in HIV and non-HIV-related serious illnesses in people who used them, leading to a widespread consensus that the strategy is best avoided. Nevertheless, researchers including the Spanish and Italian TIBET study group continue to explore scenarios in which guided treatment interruptions might be effective.

The TIBET study

The TIBET researchers, from hospitals in Spain and Italy, looked at 201 HIV-positive adults on antiretroviral therapy, with strong markers of immune health: CD4 cell counts over 500 cells/mm³ (with no nadirs below 50 cells/mm³ – the median was 338 cells/mm³ in the guided treatment interruption group) and viral loads below 50 copies/ml. (The group, enrolled between May 2001 and Jan 2002, included 27% women and a wide range of risk categories, at a median age of 38.)

Study participants were randomly assigned to either continue with standard antiretroviral therapy (control group, n=101) or to guided treatment interruptions (the GTI group, n=100). The SMART study had stopped therapy at CD4 cell counts over 350 cells mm³ and re-initiated so as to keep CD4 cell counts over 250 cells/mm³. The TIBET study used a higher threshold: participants were allowed to interrupt therapy until CD4 cell counts fell below 350 cells/mm³, HIV viral load rose above 100,000 copies/ml, an AIDS-defining illness occurred or a severe or prolonged acute retroviral syndrome (ARS) took place. In any of these cases, treatment was restarted until CD4 cell counts rose above 500 cells/mm³ and viral load fell below 50 copies/ml.

Over the two-year study, people in the GTI group stopped therapy a total of 157 times and resumed 110 times. The total median time off therapy in the GTI group was over 76 weeks – a 67% decrease in drug exposure. Most people cycled back on to therapy only once or twice; nine people went through three off/on cycles over the course of the study.

The GTI group had slightly fewer drug-related adverse events – each person had a 2% chance per year of an adverse event vs. 5.4% per person per year in the control group. However, this was more than outweighed by an increased occurrence of non-medication-related symptoms: the GTI group had much more frequent “mononucleosis-like” symptoms including fever, chills, muscle pain, and tiredness (56% per person per year, vs. 3% for controls). Ninety percent of these symptoms occurred after treatment stoppages; researchers surmised that they were part of a “retroviral syndrome” similar to that often seen after initial HIV infection.

Overall, there was no net decrease in physical symptoms in the GTI group, and patients in that group reported improvements in mental and psychosocial but not physical quality of life. (No AIDS-defining illnesses or deaths occurred in either group.) Nucleoside resistance mutations were found in 36% of 87 GTI patients, and 8 of 19 patients on non-nucleoside-based therapy (42%) showed NNRTI mutations. (Resistance was difficult to quantify in the control group since most viral loads remained undetectable.)

At any given time, 10% of the GTI group were actually below the 350 CD4 cell threshold, which “may be associated with an increased risk for clinical progression.” However, since this value was deliberately chosen to be higher than the 250 cell/mm³ cell cutoff used in the SMART study, actual clinical complications were few and minor. Researchers concluded that “GTI in their current design cannot be considered as safe as continuous antiretroviral therapy. Still, this strategy may be relatively safe in patients with nadir CD4 cell counts > 350 [cells/mm³] and pre-therapy HIV-1 RNA

Intermittent therapy – a review

An editorial review by Bernard Hirschel, one of the pioneers of treatment interruption studies, in the same issue systematically compared all major treatment interruption studies to date. Although SMART seems widely regarded as the final “nail in the coffin” for structured treatment interruptions, “of … six randomized studies, four (BASTA, ACTG 5102, HIV-NAT 001.4 and Staccato) used a CD4 cell count restart threshold of 350-400 [cells/mm³] and did not see increased serious morbidity, and two studies (Trivacan and SMART) used a CD4 cell count restart threshold of 250 [cells/mm³] and [did see an increase]. This suggests that the … restart threshold is important”. (The TIBET study could be added to the former list.)

In both SMART and TIBET, the hoped-for results (decreased cardiovascular and liver problems, increased physical quality of life) were not seen. The reviewers bluntly state that “in our opinion… ‘Once you have started HAART, you never can stop again’.” They concede that “many patients may feel that their ‘quality of life’ is improved by time off drugs, and the one in 50-100 risk of something bad happening is worth taking”, but pessimistically feel that “given the SMART results… it is unlikely that another large prospective study with clinical endpoints will be funded.”