Treatment breaks set for a come-back?

Two and half years after a major study of HIV treatment interruption was halted because of a higher risk of death in people who stopped treatment, results from a four-year Italian study reported this week suggest no elevated risk of illness or death as a result of interrupting treatment. The findings were presented this week at the Ninth Congress on Drug Therapy in HIV Infection in Glasgow.

The SMART study was a major international randomised trial of structured treatment interruption versus continuous antiretroviral therapy. Participants with CD4 counts above 350 cells/mm³ on antiretroviral treatment were randomised to interrupt treatment until their CD4 counts fell below 250 cells/mm³, or take antiretroviral therapy continuously.

The study was halted after two years because people who interrupted treatment had an increased risk of death and HIV disease progression, and subsequent analysis showed that they also had a higher risk of non-AIDS related serious adverse events including cancers and cardiovascular disease.

As a consequence of those findings, treatment interruptions, once popular, have been warned against and discarded as a potential strategy for managing long-term treatment.

So it was with some trepidation that Dr Franco Maggiolo of Bergamo, Italy, presented results from the LOTTI study, a lengthy Italian study of structured treatment interruption (STI). “My task this morning is not easy: it’s a difficult thing to talk about Strategic Treatment interruptions (STIs) since the SMART Trial, but I’m here to convince you there are some good options for patients,” he told the conference.

The LOTTI study showed that intermittent treatment of patients with CD4 counts over 350 produced clinical results equivalent to continuous treatment.

Alongside Maggiolo’s presentation, a study from Dr Cal Cohen of the Community Research Initiative in Boston, USA found that it was safe for patients to stop taking a regimen of Truvada (tenofovir+FTC) and efavirenz at weekends.

The LOTTI study is a lengthy (four years follow-up and still continuing) study in which 329 patients were randomised either to receive continuous antiretroviral therapy (ART) or to stop treatment when their CD4 counts reached 700 and to resume when they fell below 350. In addition to having a baseline CD4 count of over 700, patients had to have a nadir (lowest-ever) CD4 count of over 200 and be on stable ART with a viral load below 50 copies/ml.

Half the patients had a baseline CD4 count over 500. The average age was 40 and 73% were male, with an even spread of risk groups (37% heterosexual, 23% gay men and 39% injecting drug users). The average time participants had been on ART was six years.

The study’s primary endpoint was death from any cause, progression to AIDS or any condition requiring hospital admission. Secondary endpoints included any disease symptom not requiring admission and any grade 3 or 4 laboratory abnormality in tests.

After four years, the proportion of patients reaching the primary endpoint was 12.1% of those on STIs and 11.6% of those on continuous treatment, a non-significant difference. These results were achieved with the STI patients staying off therapy two-thirds of the time compared with less than 2% of the time in the continuous-therapy patients. Patients with CD4 nadirs exceeding 500 were able to stay off therapy 85% of the time.

After the presentation an audience member asked if Maggiolo’s study just proved that “those who should never have started can safely stop,” but Maggiolo said that only a third of participants had a CD4 nadir over 350 cells/mm³, so the majority had been eligible for ART under current guidelines.

Maggiolo detailed the clinical endpoints observed in the trial. Four patients on continuous treatment had cardiovascular disease requiring hospitalisation compared to none on STI, and six controls developed diabetes (a secondary endpoint) compared with none on STIs. In contrast, seven patients on STIs got bacterial pneumonia compared with no controls, and this was statistically significant.

There were significant differences in cardiovascular and metabolic events (primary and secondary endpoints) between controls and people on STIs: they occurred in 10.5% of controls and 1% of STIs, an incidence of 0.2% a year in STI patients compared with 3.3% a year in controls. This contrasted with the SMART study which found an incidence of 1.1% per year each in both control and intervention groups for this combined endpoint.

Maggiolo commented that the average follow-up time in SMART had been just over a year compared with four years in his study and that if the LOTTI study had been stopped at the same timepoint as SMART, no difference would have been observed in cardiovascular and metabolic events either. His study demonstrated that it took a long time for the benefits of therapy-sparing regimes to become evident.

He also said that in his study 95% of patients had CD4 counts over 350 compared with 65% on SMART and 0.5% below 250 compared with 8.6% on SMART, which might explain the higher mortality and morbidity rate seen in patients off treatment in the latter.

Maggiolo commented that doctors might like to see more evidence before making STI’s on the LOTTI model standard clinical practice, but when asked what it would take to change clinical practice commented, “It already is mine. It’s not for every clinician or every patient, but it’s a safe option for some”.

Following Maggiolo’s presentation Cal Cohen presented results from the FOTO (Five On, Two Off) study which randomised 60 largely gay male patients either to continue treatment with Truvada plus efavirenz or to stop taking it at weekends.

The patients were 83% male and 77% of white ethnicity, with an average age of 47 in controls and 42 in the intervention arm and an average CD4 count of 670.

After 24 weeks, 80% in the control arm and 83% in the intervention arm had a viral load under 50, a non-significant difference. There were five dropouts in the FOTO arm and two in the control arm. Excluding these, a ‘on treatment’ analysis shows that 100% of the intervention arm had viral loads under 50 by week 24 compared with 85% in the control arm. Most dropouts were due to the frequent monitoring required of participants, but one participant in the FOTO arm dropped out due to efavirenz-related side effects (dizziness and insomnia) and one before the study even began because he was anxious about stopping his drugs.

Eighteen patients – eight controls and ten FOTO patients - had viral load ‘blips’ during the study, defined as a single viral load between 50 and 500; the highest reading was 160 on FOTO and 465 on control.

CD4 cells at week 24 were 705 in the control arm and 635 on FOTO, but this apparent difference, Cohen commented, has adjusted since then and the latest CD4 results are the same in both arms.

Did patients adhere to the strategy? In the FOTO arm, three patients took over five doses a week at some point while eight took more than two days off, though none took more than three. No patient who took three days off developed a detectable viral load (viral loads were measured on Mondays). Unlike the Maggiolo strategy, the FOTO protocol also kept patients undetectable, meaning that it did not entail the problem of raising the infectiousness of patients.

However Cohen said he had not yet determined adherence rates in the control arm, which is obviously crucial to determining the feasibility and risks of the strategy.

Why do the strategy? Cohen asked. Apart from saving 29% of drug costs, patients much preferred it. When asked on a scale of 0-10 whether they preferred stopping their drugs at weekends, where zero indicated total disapproval and 10 total approval, the average score was 9.5.