Immediate ART associated with fewer deaths in Thailand, but starting a week or two later may lead to less virological failure

Does giving people their first HIV medications the day they are diagnosed work better than if they start them later? Or are results better if people have time to come to terms with their diagnosis? A huge study from Thailand finds it’s usually better to start in the first month – but that getting your medication the day you are diagnosed is not so clearly beneficial.

The evidence for same-day antiretroviral therapy (ART) in terms of viral suppression, mortality, and the proportion who drop out of care is ambiguous. For instance, a study from Lesotho published in 2018 found higher rates of retention in care and viral suppression in people who started ART immediately, while a 2021 study from nearby Eswatini found lower rates of retention and viral suppression.

These studies featured 441 and 1899 people, respectively. A much larger study from Thailand was presented at last month’s 12th International AIDS Society Conference on HIV Science (IAS 2023) in Brisbane. It featured all Thai adults – over a quarter of a million people – who started ART between the start of 2014, when treatment for all with HIV regardless of CD4 count was recommended, to the end of 2022. Same-day ART was recommended in national guidelines from January 2021.

Dr Sirinya Teeraananchai from Kasetsart University in Bangkok told that conference that starting ART on the same day, or at least within a week, of diagnosis was linked to a lower risk of subsequently dying. On the other hand, starting therapy that day was associated with a higher rate of disengagement from care, and was also associated with a higher risk of having a viral load above 200 after the first year on therapy.

A total of 252,239 people were included, and divided into four groups: people who started the same day or within a week of diagnosis (25%, most of whom started on the same day); people who started more than a week but less than a month after diagnosis (24%); people who started between one and three months after (23%); and people who started at least three months after (28%).

Their average age was 34 and 68% were male; these characteristics were not associated with how soon they started ART. Later year of diagnosis was associated with starting earlier. Between 2012-13 and 2020-2022 the proportion starting within a week increased from 29% to 38%, and the proportion starting more than three months later fell from 44% to 22%. Baseline CD4 count and the proportion of people who took dolutegravir rather than NNRTI-based therapy was higher in people who started ART within a week, probably because they were more likely to be diagnosed recently. Baseline CD4 count was 233 in the whole group, 319 in early ART starters and only 145 in people who started from 1-3 months after diagnosis. However, more people starting ART within a week had AIDS (28%) than those who started within a month (24%).

The lowest mortality rate was seen in people who started ART within seven days (1.28%), compared to 2.04% in people who started more than a week but less than a month later, and just under 2.5% in people who started more than a month later.

The rate of outright viral failure – defined as having a viral load over 1000 more than six months after starting ART – was also lower in early starters, but was actually lowest in people who started one week to one month after diagnosis. Starting 1-7 days after diagnosis was associated with an annual virological failure rate 0f 2.76%; starting 8 days to a month later with a rate of 2.33%; 1-3 month later with a rate of 3.35%; and more than three months later with a rate of 3.78%.

This pattern persisted: three years after starting ART, the proportion who had had viral failure during that time was 6.9% in people who started from a week to a month after diagnosis, but 8.2% in the earliest starters, and over 10% in later starters.

In multivariate analysis, the advantage of starting early, but not too early, becomes clearer. When the findings were controlled for CD4 count at presentation and year starting, the risk of virological failure was 21% lower in people starting at months 1-3 after diagnosis compared with people who started therapy more than three months after diagnosis, 38% lower in people starting a week to a month after diagnosis, but only 13% lower in the early starters.

That’s viral failure necessitating a therapy switch. But the proportion of people who had a viral load above 200 a year after starting therapy, which would not necessarily result in a switch, was actually highest in people who started within a week of diagnosis. Nineteen per cent of early starters had a viral load over 200 a year after starting, compared with only 10% in people who started a week to a month after diagnosis, and only slightly more in people who delayed further. In the latest time period (2020-22) the proportion with a viral load over 200 was 15% in the earliest starters – but only 8% in those starting within a month.

The proportion of people who dropped out of care was highest in early starters too. Thailand tends to have excellent retention rates and monitoring. Nonetheless, the proportion of early starters who dropped out of care was 2.69% in early starters, 2.47% in one-week-to-one-month starters, but only 1.51% in people who started later.

One important limitation of this study is that people are often diagnosed within the private health care system in Thailand but are then referred to public-sector HIV centres for ART. This could mean that some people's actual date of HIV diagnosis was earlier than the confirmed date recorded by the HIV centre.

The advantage of immediate ART in terms of mortality is hard to argue with, though it could be explained if immediate ART was more likely to be offered to people who presented with AIDS-defining illness. The study's other findings suggest that people who have had a week or more to think about it and then start may do best of all.