Planned versus unplanned interruptions
CD4 count-guided treatment interruptions as a long-term HIV treatment strategy have been extensively discussed in recent months, after the early discontinuation of the SMART study in January. The SMART investigators found that people who were randomised to interrupt ART once CD4 counts rose above 350 cells/ mm3 and restart once CD4 counts fell below 250 cells/ mm3 had an increased risk of disease progression and death compared to those randomised to remain on ART. Consequently, planned treatment interruptions are not currently recommended by treatment guidelines.
Nevertheless, many HIV-positive individuals continue to interrupt ART for a variety of reasons. These unplanned interruptions often occur due to toxicity, treatment fatigue, and situations where the treatment of another infection - such as TB - might jeopardise the effectiveness of either treatment. In addition, a variety of social and mental health factors may also lead to unplanned interruptions.
One-in-six previously treatment-naïve individuals were likely to interrupt treatment for at least six months within two years of starting antiretroviral therapy (ART), according to the results of a large observational study published in the August 15th edition of the Journal of AIDS. The study also found that discontinuing treatment appeared to be safe over the short-term for most people, although the risk was higher for people aged over 40, those with a pre-treatment CD4 count below 200 cells/mm3 or with a CD4 count below 350 cells/mm3 prior to interrupting therapy.
How many interrupted or switched treatment?
The participants took their first ART combination for a median of 19.9 months. In total, 299 (19.3%) interrupted therapy (defined as a discontinuation of all drugs for at least 14 days), and another 283 (18.2%) switched to a new ART combination during the study period. The remaining 969 (62.5%) neither interrupted nor switched throughout their follow-up.
The investigators calculated that the cumulative probability of interrupting treatment two years after starting ART was 15.9%, and the cumulative probability of switching to a new regimen two years after starting ART was 16.3%. This cumulative probability increased to 29.7% and 29.1%, respectively, within five years of initiating ART.
Multivariate analysis found that women (who acquired HIV via either sex between men and women, or via IDU) were more likely to interrupt their treatment (hazard ratio [HR] 1.62; p=0.005).
The investigators say that reason for this gender difference “is unclear…[although] recent studies reported lower ART adherence rates in women compared with men with this difference being explained mainly by social and behavioural factors." An alternative suggestion is that women may experience more drug-related side-effects.
Additionally, IDUs were both more likely to interrupt treatment (HR 1.55; p=0.005) and less likely to switch to a new regimen (HR 0.56; p=0.001). The investigators suggest that this is most likely due to adherence issues.
Other factors associated with interrupting treatment include a higher viral load at baseline, a higher CD4 count at baseline and a poor CD4 response to ART, which, the investigators write “are in line with results from previous studies.”
Of interest, the investigators found that time between seroconversion and initiating ART, or the rate of CD4 decline prior to starting does was not associated with interrupting treatment.
CASCADE collaboration
Investigators from the Concerted Action on Seroconversion to AIDS and Death in Europe (CASCADE) collaboration - which pools data from 22 cohorts comprising 8300 HIV seroconverters from Europe, Australia and Canada - sought to ascertain the incidence and predictors of treatment interruptions in clinical practice; and to assess CD4 cell count and viral load changes during these unplanned interruptions, as well as risk factors for these changes.
One of the strengths of CASCADE is that all participants have a well-estimated seroconversion date, which enabled the investigators to assess the effect of participants' pre-ART history on the consequences of treatment interruptions.
However, a major weakness of this study is that no data were collected that provides information regarding the reasons for interrupting or changing therapy, or on adherence to ART. A recent audit by the British HIV Association found that lack of effectiveness of ART was a reason to switch from first- to second-line therapy in only a third of cases, whereas toxicity or side-effects (affecting more than 50%, and not measured in the CASCADE study of treatment interruption) was the main reason for switching.
Of the 8300 participants in CASCADE, 1551 met the inclusion criteria for this analysis. These included being on a stable first ART regimen consisting of at least three anti-HIV drugs for at least one year after seroconversion; and having baseline and at least one additional post-ART viral load and CD4 cell count measurement.
The majority of the 1551 participants were male (83%), and the median age at ART initiation was 35.9 years (the median age at seroconversion was 28.5 years). Just over half of the cohort (56%) were infected with HIV during sex between men in 56%; 22.5% were infected through injecting drug use (IDU) and 15.7% were infected during sex between men and women.
Time off treatment and clinical progression
The median duration of treatment interruption for the 299 participants who took an unplanned treatment break was 199 days, or just over six months. However some interrupted their treatment for much shorter or longer periods: the interquartile range was 103 to 416 days.
The 299 individuals who interrupted ART were followed-up for 472.8 days on ART and accounted for 242.8 person-years of follow-up during treatment interruption. The 969 individuals who remained on ART throughout the study accounted for 3115 person-years of follow-up.
The investigators only measured clinical progression via the development of a clinical AIDS event, or death. Seven people who did not have AIDS when they began ART developed a clinical AIDS event during their treatment interruption. Eight people who did not have AIDS when they began ART developed a clinical AIDS event whilst on ART. In addition, two people with AIDS when they began ART experienced another AIDS illness during their treatment interruption, compared with one person with prior AIDS who remained on ART.
However, the investigators found that there was no statistically significant difference between the rate of developing a new AIDS illness whilst interrupting treatment compared with remaining on ART. (0.037 vs. 0.019 cases per person per year; p=0.158).
Although the investigators note that nine individuals in the study died without an AIDS diagnosis, they do not provide any further data. Since the SMART study found an increased risk of non-AIDS deaths during treatment interruptions, information on whether these deaths occurred on or off treatment would have been illuminating.
Immunological and virological consequences of interruption
Prior to interrupting treatment, the median CD4 cell count was 441 cells/mm3. The median CD4 cell count drop during interruption was 94 cells/mm3. They calculated that this was the equivalent of losing, on average, 57.3% of an individuals' CD4 cells each year off ART.
However, the investigators found that individuals who began ART aged over 40, or who had a lowest-ever CD4 cell count that was below 200 cells/ mm3, and/or who had a CD4 cell count that was below 350 cells/ mm3 when they stopped treatment, had the fastest declines in CD4 counts off treatment.
Of note, the rate of CD4 cell decline before ART initiation was not significantly associated with the rate of CD4 cell decline when interrupting treatment.
The median viral load of those who interrupted treatment was less than 500 copies/ml when they interrupted ART. This increased rapidly during the first three months off treatment (by a median of 1.47 log10 copies) but then became relatively stable.
Conclusion
The investigators conclude that “a substantial and increasing number" of individuals interrupt their treatment during their first few years on ART, but add that this “seems to be safe for [individuals] with a well-retained immune system.”
They caution that people older than 40 years old; individuals who have ever had a CD4 count below 200 cells/mm3 “or with limited immune reconstitution” whilst on ART, “had the greatest proportionate decrease in CD4 cell counts” during an unplanned treatment interruption.
They recommend that for these people in particular “caution and close monitoring are essential to ensure that risks are minimal.”
Touloumi, G et al. Highly Active Antiretroviral Therapy Interruption: Predictors and Virological and Immunologic Consequences. Journal of Acquired Immune Deficiency Syndromes 42(5): 554-561, 2006.