The cornerstone of HIV treatment may be losing some of its power: will dolutegravir resistance become a problem?

A number of presentations at last week’s Conference on Retroviruses and Opportunistic Infections (CROI 2024) looked at whether HIV is starting to develop resistance to dolutegravir, one of the most widely used antiretrovirals in the world.

Guidelines in most parts of the world recommend dolutegravir (or its close cousin, bictegravir) as a component of most people’s first HIV treatment regimen. The World Health Organization recommends it for all populations, and if it wasn’t part of a person’s first-line HIV treatment, it will usually be part of their second-line regimen. The majority of low- and middle-income countries have now switched to a combination pill of tenofovir, lamivudine and dolutegravir (TLD), including South Africa, home of the world’s largest antiretroviral therapy (ART) programme.

This was possible due to dolutearavir manufacturers ViiV allowing generic manufacturers to make it, but dolutegravir was chosen in the first place because studies showed that it was both pretty free of side effects – after an early scare about birth defects was disproved – and because dolutegravir appears to pose a very high barrier to the development of HIV drug resistance.

However, dolutegravir is now such a keystone of ART worldwide that programmes need to keep a careful watch for any sign of an increase in dolutegravir-resistant HIV. Last week, the World Health Organization issued a report raising concerns that recent observational surveys have found that “levels of HIV dolutegravir resistance are exceeding levels observed in clinical trials”. The next day, an entire poster discussion session was devoted to the issue at CROI 2024.

Introduction

Dr Gert van Zyl of Stellenbosch University in South Africa introduced the session by commenting how dolutegravir had become the “one drug to rule them all”. He said the session had been prompted by reports of dolutegravir resistance in studies such as this one, mainly from high-income countries and with clade B virus, where 14% of people with ART failure had integrase strand transfer inhibitor (INSTI) resistance mutations, though only 1% had high-level resistance to dolutegravir.

He said that the problem could in theory be worse in lower-income settings where far more people have been switched to dolutegravir. Some were switched as long as six years ago, often regardless of whether they were virally suppressed at the time they switched, and in the absence of ongoing viral load monitoring. In addition, non-B clades of HIV might have different integrase resistance patterns.

It was not easy, he added, to judge the degree to which resistance was a problem. There is a problem of which denominator was used when reporting the prevalence of dolutegravir resistance – was it everyone on the drug, only people with virological failure, or only ones suspected of having resistance? Then there was also a problem of what numerator was used – was any dolutegravir mutation counted (even with no loss of susceptibility), only mutations classed as high level, or only cases where there was clear loss of susceptibility?

Van Zyl listed the resistance mutations most commonly seen so far, as collected by the Stanford University Resistance Database. Among eight regarded as ‘major’, the ones seen most often in the studies presented in this session were changes to the amino acids in HIV’s integrase enzyme at positions 66, 118, 138, 148 and 263. Several studies also saw a mutation at position 147, not in Stanford’s list.

These ‘phone numbers’ of mutations are worth remembering not so much in themselves as because they seem to occur in very different patterns in different studies, and seem to be influenced by whether people are taking INSTIs for the first time or not, and whether they have other mutations, even to other drug classes.

In general, significant dolutegravir resistance still seems to be uncommon, and can even be reversed with improved adherence. But it has got more common in the last two years, and certain groups, including children, seem more prone to it.

Dolutegravir resistance in children

One study exclusively concerned children in Malawi. Dr George Bello of the International Training and Education Center for Health in Lilongwe, Malawi, told the conference that to be included in the study, children had to be aged 2-14, to have been on dolutegravir for more than nine months, and to have had a latest viral load of over 1000, which was still over 1000 when tested again after adherence support for child and parent.

The sample represents a very small proportion of children with HIV in Malawi. At the latest count there are 53,843 children with HIV, of whom two-thirds (35,352) are on ART, 99% on dolutegravir. This study enrolled 302 children attending 19 of the 25 largest HIV clinics in Malawi. After adherence counselling, 169 (56%) achieved viral suppression again.

The remaining 133 children had tests for drug resistance to every antiretroviral (ARV) drug class. Three-quarters were found to have at least one resistance mutation, but resistance to both classes of reverse transcriptase inhibitors was far more common, with 65.5% having resistance to non-nucleoside reverse transcriptase inhibitors (NNRTIs) and 42% to nucleoside reverse transcriptase inhibitors (NRTIs). In contrast, only 16% had significant INSTI mutations. Protease inhibitor (PI) mutations were even rarer at 5%.

Of 18 children with successfully sequenced INSTI resistance mutations, two quite separate patterns emerged. Eleven children had the resistance mutation at position 263 and five had one at position 138. Only one child had neither of these mutations and in nine cases the 263 mutation was the only one. We’ll encounter these mutations again.

Eighteen children out of more than 35,300 on dolutegravir cannot be the sum total with high-level resistance; but it does indicate that, while dolutegravir resistance is still uncommon, compared especially with the legacy of years of NNRTI-based regimens, it is not insignificant.

A second study hinting that children might be more at risk than adults came from Lesotho. This study by Professor Niklaus Labhardt of the University of Basel in Switzerland was of a cohort in the two easternmost provinces of the country. This cohort includes 17,724 members but under 3% of them are children.

To be tested for resistance, cohort members had to have changed from NNRTI-based regimens to TLD, and to have had two successive viral loads below 50 or one over 500, with the last taken more than 16 months after switching.

Of 15,299 who switched, 151 were eligible and 78 had a resistance test. Of these, only eight (10%) had dolutegravir resistance, but two of these were boys aged nine and seven – 25% of those with dolutegravir resistance in a cohort of whom less than 3% are children. The association with age did not stay statistically significant in multivariate analysis because there were so few children, and the only other characteristic, which remained significant, was to have been treated in a nurse-led local centre rather than one of the three regional hospitals.

Two more African studies

Two more studies came from adults in Africa. Kenya was one of the first countries to roll out TLD, in 2017. Dr Leonard Kingwara of the Kenyan Health Ministry tested samples from 55 people who were on TLD and had viral loads over 200. Forty-four of the 55 samples returned a valid resistance result (it is often difficult to test samples with low viral loads). Of the 44, eight (14.5%) had dolutegravir resistance mutations. Only one of the 12 who were new to ART when they started TLD had resistance, but, seven of the 31 (nearly a quarter) who had been on previous NNRTI-based regimens had resistance.

The same pattern of two different resistance mutation patterns was seen here as in the Malawi study. All seven of the people with prior NNRTI experience had a handful of mutations that included the one at position 138 (and usually at least one other). In contrast the one previously ART-naïve person with dolutegravir resistance was the only one with a mutation at position 263, and only at that position.

In contrast, a study from Zambia and Malawi found only two people with significant dolutegravir resistance among 2833 people who had switched from TLD.

Dr Richard Lessells of the KwaZulu Natal University said that two years after the switch they only found 92 people with viral loads over 400, representing 4.7% of the people in Malawi and 1.8% in Zambia. The difference was due to people with detectable viral loads in Zambia not being eligible for switching, unlike in Malawi, and this was the one clear risk factor for virological failure, increasing the odds of failure sevenfold.

Sixty-two people had a viral load over 1000 and 45 samples were successfully sequenced. Of these, only two had high-level dolutegravir resistance. In both cases the resistance mutation was at position 138 along with at least one other major mutation, but not the 263 mutation.

Emergent dolutegravir and bictegravir resistance in Mexico

Across the Atlantic, a study of integrase inhibitor resistance included both of the second-line INSTIs, dolutegravir and bictegravir. Dr Luis Soto-Ramirez of the Salvador Zubiran National Medical and Nutrition Science Institute in Mexico City told the conference that about 87,000 of the 340,000 people living with HIV in Mexico are now taking regimens based on dolutegravir or bictegravir. Many have switched, not from NNRTIs or PIs, but from regimens based on the two earlier INSTIs, raltegravir and elvitegravir.

The institute was able to gather 100 resistance tests from people who had experienced virological failure on bictegravir or dolutegravir. Of these, 75 had previously also had treatment failure on the first-line INSTIs. The team managed to successfully sequence 79 of the 100 samples for resistance mutations and found 20 with primary mutations.

As in other studies, INSTI resistance fell into two patterns; resistance with the mutation at position 263 and resistance without it. In the seven samples with the 263 mutation, four were from people on bictegravir. Of these, all had an additional mutation at position 50 which is not counted as a resistance mutation but instead reduces the viral load. The average viral load in the people with this ‘M50I’ mutation was 2250. All but one of the people on bictegravir had only these two mutations.

In contrast the three people on dolutegravir had five, six and nine mutations respectively. The person with nine mutations also had M50I, resulting in a low viral load, but the people with six and five mutations had viral loads of 35,000 and 300,000 respectively. None of the people with the 263 mutation had taken another INSTI.

The other pattern was a group of seven people who all had mutations at positions 140 and 148 and four at position 138 too. These mutations are the ones seen as conferring the highest-level resistance to dolutegravir. Four of the seven had also taken the first-generation INSTI raltegravir.

While the frequency of this pattern did not appear to be on the increase, the 263 mutation has become more frequent. This pattern was seen in just 0.3% of resistance tests carried out for a previous study in 2019-2021, 5% in samples collected in 2021-22, and 10% in 2023. This mutation seems therefore to be a specific signal of first-time INSTI resistance to either dolutegravir or bictegravir – and may possibly be less damaging if to the latter.

Another interesting fact was that the NRTI mutations of 11 people with INSTI resistance were sequenced too, and not one of them turned out to have the M184V/I resistance mutation to lamivudine / emtricitabine, despite this being the most common NRTI mutation, and despite most of the 11 having other mutations. So far there’s no explanation for this, though M184V is known to lower viral load and could therefore make it harder for HIV to jump over the high resistance barrier to INSTIs.

What comes after dolutegravir?

Finally, if people do experience treatment failure with the second line INSTIs, what treatment options do they have? A study from Dr Ying Zhao of the University of Cape Town looked at people whose second-line protease inhibitor-based regimens had failed – this was usually boosted lopinavir. Of 355 people, 234 had resistance mutations to protease inhibitors. Of these, 133 switched to dolutegravir, compared with 101 who switched to boosted darunavir, regarded as a third-line PI. People who switched to TLD were also compared to another 121 who had failed second-line PIs but did not have PI resistance mutations.

Staying on the old PI did not work well; after 12 months only 35% had suppressed virus, despite not having resistance at baseline, compared with 77% who switched to TLD.

However, switching to boosted darunavir worked just as well as switching to dolutegravir. Eighty-nine per cent and 92% of people on dolutegravir and darunavir respectively maintained viral undetectability at 12 months, despite having PI resistance mutations. This would appear to show potential for boosted darunavir to be a salvage therapy for the still thankfully few people whose second-line INSTI regimens stop working,