Time spent by people in HIV care in US with a transmissible viral load has fallen by three-quarters since 2000

A study by the US Centers for Disease Control and Prevention (CDC) presented at the recent Conference on Retroviruses and Opportunistic Infections (CROI 2017) in Seattle found that the proportion of time people with HIV spend in care but not virally suppressed has fallen from 40% to 10% in the last 15 years.

It also found that young people, black people and people with public rather than private health insurance spent less time virally suppressed.

The biggest problem the United States faces in reaching the UNAIDS 90-90-90 target and having 72.9% of its entire HIV-positive population virally suppressed is the high proportion in that country who are diagnosed but not in care. It is estimated that 61% of HIV transmissions in the US come from people in this situation.

Some transmissions, however, still come from people who are in care but are not virally suppressed (the ‘third 90’).

The CDC’s Kate Buchacz analysed figures from the HOPS cohort, an open, prospective cohort of 5000 HIV-positive people in care at a varied selection of HIV clinics, mainly hospital-based, in nine US cities. HOPS has been established since 1993 but this study looked at viral suppression in HOPS patients from 2000 to 2014.

The definition of being ‘virally suppressed’ was the proportion of time spent with a viral load below 1500 copies/ml, which is not the usual definition of ‘undetectable' (usually below 50 copies/ml), but is the one used by the World Health Organization as the threshold for infectiousness, as very few transmissions have ever been seen from someone with a viral load below this figure.

There were 5873 people in the study with an average follow-up of 5.4 years, amounting to 37,794 person-years of observation altogether. On average 15 viral load tests were taken per person with a median time of 3.6 months between tests.

During that time people spent 86% of their time on antiretroviral therapy (ART) and 14% off it. The amount of time people spent on ART but not virally suppressed was 13% over that whole time period. In addition there were 4% of people who were not on ART but spent some time with viral loads under 1500 copies/ml.

The amount of time people spent not virally suppressed during one year fell over time. In 2000 36% of people’s time on ART was spent with viral loads below 1500 copies/ml and was as high as 40% in 2003. It then fell steadily to 10% in 2014.

However, this includes time people were not on ART: while people in HOPS spent 90% of their time on ART in 2000, this fell to 80% in 2003 to 2004, then rose steadily to 93% in 2014. Not all the people off ART were drug-naïve: in 2003-2004 more than half of those not spending time off ART (11% of people) were treatment-experienced and taking breaks. By 2014, only 3% of patient time was spent having a break off treatment. This may have coincided with the period of maximum concern about long-term toxicities: the SMART study, the first one to show that it was generally better to be on ART than not, announced its results in 2006.

The proportion of time people spent not virally suppressed while on ART and therefore in true treatment failure or yet to be suppressed was 31% in 2000, falling to 7% in 2014. This varied somewhat by drug class. In 2000, 22% of patient-time on NNRTI-based regimens was spent with viral loads over 1500 copies/ml versus 32% of time on protease inhibitor (PI) regimens: this proportion of patient-time had fallen to 4% and 11% respectively by 2014. This does not take account of the fact that people on PI regimens are more likely to be on second-line therapy and/or have drug resistance. In 2014 9% of patient-time of people on integrase inhibitors, which were unavailable in 2000, was spent with viral loads over 1500 copies/ml.

In multivariate analyses, certain groups of people were more or less likely to spend time not virally suppressed. Older people were less likely, with the amount of time virally suppressed rising by 9% for every ten years older, and people aged below 35 on ART were 50% more likely to spend time with viral loads over 1500 copies/ml than people over 50. People with public rather than private health insurance were 24% more likely to spend time with viral loads over 1500 copies/ml, and black people 19% more likely than white people. Women were more likely than men to have periods with viral loads over 1500 copies/ml – 31% versus 22%, or 26% in heterosexual men alone – but this association lost significance after controlling for race and insurance status, i.e. women were less likely to spend time unsuppressed because they were black or had public insurance, not because they were women.

This study reinforces the findings from another presented at the conference, which showed that disparities in age, race and insurance status continue to affect the success of antiretroviral treatment. As this study was of people in care who were largely on treatment, it shows that access to treatment may not be the sole determinant of viral load; adherence and drop-outs from treatment caused by insurance problems may also have a role to play, though HOPS did not directly measure adherence.