Community perspective on the INSIGHT Strategic Timing of AntiRetroviral Treatment trial

Abstract

Determining when to start antiretroviral treatment (ART) is vitally important for people living with HIV. Yet the optimal point to start to maximise clinical benefit remains unknown. In the absence of randomised studies, current guidelines rely on conflicting observational data and expert opinion, and consequently diverge on this point. In the United States, ART is recommended irrespective of CD4 cell count. The World Health Organization now recommends starting ART at a CD4 cell count of 500 cells/µL, while the threshold for the United Kingdom and South Africa remains at 350 cells/µL.

The Strategic Timing of AntiRetroviral Treatment (START) study, one of the largest clinical trials on the treatment of HIV, will answer this question. START compares two treatment strategies: immediate treatment at a CD4 cell count of 500 cells/µL or higher versus deferring until the CD4 cell count decreases to 350 cells/µL or AIDS. START includes seven substudies, five of which will clarify the relative contribution of HIV and ART in common comorbidities. START is fully enrolled and expected to be completed in 2016. HIV advocates support the study’s design and have been involved from inception through enrolment. The trial will produce rigorous data on the benefits and risks of earlier treatment. It will inform policy and treatment advocacy globally, benefitting the health of HIV-positive people.

Introduction

UNAIDS estimated in 2013 that more than ten million people are on antiretroviral treatment (ART) worldwide and has called for this to increase to fifteen million people by 2015 (1, 2). The effect on public health has been profound. Studies suggest that life expectancy is normal for people with uncomplicated HIV infection with access to modern treatment and care (3). In South Africa, life expectancy increased by seven years between 2005 and 2013, primarily due to large-scale ART rollout (4). However, globally, and in all settings, late diagnosis continues to drive mortality, morbidity and, possibly, incidence.

Despite this progress, the optimal CD4 cell count at which to begin treatment remains unknown. Identifying the optimal time of initiation can improve treatment outcomes. Data showing the safety of initiating ART earlier may drive earlier testing and treatment and reduce stigma.

Rationale for START

The Strategic Timing of AntiRetroviral Treatment (START) trial will help answer this question (5). It is a study that has been driven by community demand that the optimal clinical initiation threshold for ART be determined by clinical trial evidence rather than expert opinion informed primarily by observational data (6).

This open-label multinational trial has randomised participants with CD4 cell counts greater than 500 cells/µL either to begin treatment immediately or to defer ART until the CD4 cell count has declined to 350 cells/µL or until AIDS symptoms develop (5). It is planned to complete by the end of 2016 when the anticipated number of primary endpoints reaches 213 (7). With 4685 participants, it is the largest ongoing ART trial.

Current HIV treatment guidelines vary considerably on when to start treatment and reliance on expert opinion in the past has often underestimated the risks of treatment and overestimated the benefits of earlier initiation of ART. For example, in 1987 the U.S. Department of Health and Human Services (DHHS) guidelines and others set the threshold at 500 cells/µL, where it remained for 14 years, during suboptimal treatment with monotherapy, dual therapy and early triple combination therapy. It wasn’t until April 2001 with accumulating data on the toxicity of the first triple combinations and the risk of drug resistance, together with a low risk of HIV-related complications at high CD4 cell counts, that the CD4 threshold was reduced to 350 and then 200 cells/µL two years later. A better understanding of adherence, resistance and new concerns about being off treatment, together with the availability of better drugs, provided sufficient evidence for the DHHS to increase the threshold back to 350 cells/µL in 2007 and to 500 cells/µL in 2009. However, when the CD4 threshold was removed as a criterion for deciding when to begin treatment in 2013, this was based on low-quality evidence and expert opinion (8).

It is notable that when the World Health Organization (WHO) moved the CD4 initiation threshold from 350 to 500 cells/µL in 2013, it was for programmatic reasons and the potential to reduce further transmission, rather than based on new evidence of individual clinical benefit at higher CD4 counts (9). The threshold in the United Kingdom and South Africa remains 350 cells/µL (10, 11).

Evidence from randomised clinical trials shows that the threshold must be set higher than 250 cells/µL (12–14). Data become less clear at levels above 350 cells/µL (15–18). An analysis in 2010 concluded that there is no convincing evidence “that deferral of initiation of ART to a CD4 count of 350 per µL causes net harm, particularly in terms of mortality, compared with starting at any higher level” (19). The absence of any subsequent randomised data continues to support equipoise for both strategies in START.

Between 2009 and 2013, when START was enrolling, the attitude to ART changed in many settings and for reasons other than clinical benefit for the individual patient.

HPTN 052, a randomised trial in serodifferent couples, showed that ART significantly reduced the risk of sexual transmission of HIV and created an impetus to initiate patients earlier (14). Clinically, it highlighted the risk of waiting to initiate treatment until a CD4 cell count of 250 cells/µL, which was already lower than guidelines advised. From a public health perspective, based on these data, a universal test and treat approach would reduce incidence and, according to several mathematical models, eventually eradicate the epidemic in South Africa earlier than a 350 cells/µL threshold (20).

The observational PARTNER study of HIV serodifferent couples not using condoms when the positive partner had an undetectable viral load on treatment produced a significantly larger dataset to quantify the dramatic impact of ART on sexual transmission (21). However, this use of ART is already included in US, UK and WHO guidelines as a reason for starting treatment at any CD4 cell count, including above 500 cells/µL.

All medicines carry a risk of toxicity and drug resistance. Even if ART has a small risk of harm, the risk of serious HIV-related events at CD4 cell counts > 500 cells/µL is small, so it is possible the harm of early treatment outweighs the benefits. We have argued that the decision to start treatment must be taken by patient and clinician in consultation and should be informed by what is best clinically for the individual patient, including the potential to reduce anxiety and stigma relating to the risk of onward transmission (6). Data to inform clearly such decisions can only come from START and similar studies.

START Substudies

Irrespective of whether START finds that the benefits of earlier treatment outweigh the risks based on primary events, the considerable dataset, including from the seven substudies, will be a unique resource to inform key aspects of living with HIV over the long term.

Observational studies report higher rates of comorbidities, such as cognitive decline, cardiovascular disease, lung infections and bone disease, in HIV-positive cohorts than the general population, even in cohorts on treatment with suppressed viral loads (22).

However, irrespective of individual drugs used, ART has been associated with increased loss of bone mineral density compared to not being on treatment. There is concern that HIV continues to replicate in the brain and spinal fluid despite ART. Data on these reservoirs and the potential toxicity of ART in a specific reservoir, especially in asymptomatic persons, are lacking.

In the SMART study, which tested structured interruptions of ART for paricipants with a CD4 cell count over 350 cells/µL, cardiac events were one of the drivers of higher morbidity and mortality in the interruption arm (23). Several inflammatory markers have also been associated with higher mortality in SMART (24, 25).

START substudies will provide data that help address these and other concerns. For example, one substudy is looking at changes in arterial elasticity, and a pulmonary substudy is examining rates of lung function decline and chronic obstructive pulmonary disease.

In addition to substudies examining comorbidities, START included a substudy comparing participant understanding of informed consent forms. The results of this substudy have the potential to influence an aspect of research that many researchers think drifted from being patient-centred many years ago.

Current Status of START

START is now fully recruited. Women comprise 27% of the participants. Nearly a third of participants smoke.

When the study was planned, it was expected to appeal to two distinct groups: those recently diagnosed and still in relatively early infection and those who had been diagnosed for many years but whose immune response maintained a high CD4 cell count. Baseline data now show that the former group predominates. The median time from diagnosis was one year (IQR: 0.4–3.0) and 23% of participants had been diagnosed within four months before entering START, showing that many people found the option of early treatment an acceptable one.

The median viral load was 4.1 log₁₀ copies/mL (IQR: 3.5–4.6) showing that for most people a high rate of early viral suppression could be expected. In contrast, about 10% of participants had viral loads greater than 100,000 copies/mL and 28% had nadir CD4 cell counts less than 500 cells/µL showing a willingness to defer treatment in those at relatively higher risk of HIV progression.

It will be important in many developing countries to know whether earlier treatment will reduce rates of tuberculosis (TB). While not designed or powered to detect the effect of earlier treatment on rates of active TB, 22% of participants are from high TB areas. START might therefore provide useful data on the effect of treatment on TB. This is a question also being examined in the TEMPRANO study in Côte d'Ivoire (26).

The CD4 cell count at study entry is higher than the estimated value used in the original sample size calculation. This enables START to answer the primary hypothesis with fewer clinical events than originally calculated. The concern now is to encourage participants to remain adherent to their randomised arm until the study closes in order to maintain the power of the results.

This takes a certain degree of courage and determination from both researchers and participants. The contribution from a few thousand people, who are generally at low risk of HIV-related complications, to produce results that will inform the treatment of millions of people globally, is considerable.

Although programmes focused on “universal treatment,” “test and treat,” and “treatment as prevention” have vocal proponents with admirable goals for public health, they are driving use of treatment in advance of clear evidence demonstrating that clinical benefits outweigh the risks for the individual taking treatment. The first decades of treatment guidelines and onward to the unexpected results in SMART, highlight the pitfalls of relying on expert opinion without accompanying clinical trial evidence.

The request by HIV-positive people and our communities to have this data is a modest one. START’s successful completion will produce rigorous data on both the benefits and risks of earlier treatment to inform government policy, treatment advocacy and the health of HIV-positive people globally.