Injectable HIV treatment underway in England and largely effective, but a couple of breakthrough cases raise concern

Injectable HIV antiretroviral therapy (ART) is starting to be provided in England with several hundred patients now on the two-monthly injections. The British HIV Association (BHIVA) Spring Conference in Birmingham last week heard some of the first data from the rollout in England. In general, the injections are effective and liked by patients, though there have been a few withdrawals.

However, there have been two cases reported so far of virological failure with drug resistance. This echoes the pattern seen in the scientific trials of injectable ART, where there were 24 cases of virological failure among 2313 study participants.

This 1% failure rate would be seen as an excellent result in an oral therapy drug, but is causing concern because researchers do not yet understand how it can happen despite apparently perfect adherence, and because three-quarters of the participants experiencing drug failure have acquired resistance to one or both of the drug types used, which can limit future treatment choices.

The injectable ART uses long-acting formulations of the drugs cabotegravir (Vocabria) and rilpivirine (Rekambys). (In some other countries, including the US, the same drugs are packaged together with the brand name of Cabenuva.) They are given as two injections, one of each drug. The first two doses are a month apart and then injections are given every two months. There is an optional ‘lead-in’ of a month’s worth of tablets beforehand. This is to rule out rare cases of intolerance to cabotegravir or rilpivirine, but also because it takes time in some people for levels of the injectable therapies to build up to saturation levels in tissues.

Success and failure in UK patients on injectables in 12 clinics

Dr Kyle Ring from Queen Mary University in London presented data from 12 UK clinics – seven in London and one each in Liverpool, Cardiff, Walsall, Blackpool and Newcastle. These clinics have initiated a patient cohort called SHARE LAI-net, as part of the SHARE consortium which supports HIV and related research aiming to improve health inequalities.

So far, 518 people in SHARE LAI-net have been approved for injectable ART and 433 have received at least one injection – the others are waiting to start or are in their month of oral lead-in. To date, the average time on injectables for those who have started is 7.5 months – in other words they are on their fourth or fifth injection.

The average age of people starting injectables is 46. Thirty per cent are women including three trans women; the rest are cisgender men. Nearly half are White and 37% are Black.

Before starting injectables, 52% were on an integrase inhibitor-based regimen, 36% on an NNRTI-based regimen (including 14% on oral rilpivirine) and 13% on a protease inhibitor-based one. Eighty-four per cent were on triple therapy regimens that also included NRTI drugs; the remainder (85 people) were almost entirely on two-drug combination pills, mainly Dovato or Juluca, which contain the other second-generation integrase inhibitor, dolutegravir.

Physicians were asked for the main reasons their patients gave for wanting injectables. The most common was the “inconvenience” of daily pills, cited by 74%, but 25% said “stigma” and 21% said “fear of disclosure”. Fifteen per cent gave frequent travel as a reason.

Ninety-seven per cent of the patients receiving more than one injection have had their repeat injections on time, within the seven-day window allowed.

Twenty-five patients (6%) have discontinued their injections. Four were because of detectable viral loads, detailed below. Fourteen were due to side effects including injection site reactions, two found the injections inconvenient, two were lost to follow-up, one did not start because of non-suppression on their oral regimen, one was because of pregnancy and there was one death (not related to HIV).

Viral blips and viral breakthrough

Viral load tests are taken at every injection. Results over 50 copies/ml are classified as:

• ‘Blips’: these are single viral loads of 50-200 which are not preceded or followed by another. These are not a criterion for stopping injectables and in fact there were 24 blips in 21 people (5% of the cohort). Blips are not uncommon in oral therapy, where single blips are not usually associated with subsequent failure.
• ‘Low level viraemia’: this is more than one consecutive blip which does not exceed 200. It is not counted as drug failure, but one patient in the cohort with persistent viral loads in the 50-200 range chose to go back on oral therapy.
• ‘Viral failure’: this means more than one consecutive viral load over 200. The patient is advised to to switch back to oral therapy.

There were three cases of viral failure in the SHARE LAI-net cohort. None of them came completely out of the blue; two were immediately preceded by viral loads above 50 and the third was in someone who had had two previous viral loads above 50, including at his first injection when ending his oral lead-in phase.

The first two did not acquire drug resistance. One was a 45-year-old Black woman, who had a viral load of 75 when tested at her fourth injection. When she had a repeat test her viral load was 278, above the criteria for stopping, and it was 479 at a subsequent test. She resumed oral therapy on Biktarvy (bictegravir, tenofovir alafenamide, emtricitabine) and re-suppressed her viral load.

The second was a 63-year-old White man. He had a potential risk factor in that he was considerably overweight, with a BMI of 40, on the border of severe obesity. Obesity can cause drug distribution problems in the body and may also require the use of longer needles. His viral load was 437 at his third injection. Although it was lower on subsequent testing at 271 he was still above the criteria for viral failure so oral therapy was resumed on Delstrigo (doravirine, tenofovir disoproxil, lamivudine) and he also re-suppressed his viral load.

The third case was more worrying. This was a 40-year-old White man who turned out to have a shockingly high viral load of 109,000 when he attended for his seventh injection, 11 months after starting. A repeat viral load test was 10,300. His viral load at his sixth injection had been below 50. A resistance test showed that he had HIV with a mutation called K101E, which under normal circumstances confers roughly twofold resistance against rilpivirine – in other words, double the usual amount of drug is needed for viral suppression.

Despite the previous undetectable viral load, his viral failure did not come completely out of the blue. He had had a viral load of 64 at the end of his oral lead-in, before his first injection. As this was not subsequently repeated it was counted as a blip and he was cleared to receive the injectables.

However, his viral load on his fifth injection was 194 – very close to the limit that would have mandated a return to oral therapy. Was this indicating that he had already developed a degree of resistance to one of his drugs by his fifth injection? Or was it the first sign that for some reason his drug levels were, at least intermittently, too low? We don’t know. And we also don’t know why his viral load went up so fast within a month or why – given this very high viral load on drugs that failed to suppress it – he did not also acquire resistance to cabotegravir.

In fact resistance tests are often hard to perform, so in case he was harbouring undetected cabotegravir resistance, he was put on an oral regimen of Symtuza (boosted darunavir, tenofovir alafenamide and emtricitabine) which, as a protease inhibitor-based combination, avoids both integrase inhibitors and NNRTIs. This successfully suppressed his viral load to below 50 on this.

Another breakthrough case in Leeds

 This was not the only case of unexpected viral load breakthrough with drug resistance reported at the conference. Dr Katie Drury of the Leeds Teaching Hospitals NHS Trust, which is not in SHARE LAI-net, outlined their experience with their first 25 patients on Vocabria and Rekambys.

The hospital had in fact assessed 40 patients as eligible for injectable ART. Of these, physicians’ assessment of the main reason for eligibility was described as ‘psychological’ in 26 cases (65%) – in other words these were people who described depression, anxiety or feelings of stigma in relation to taking oral ART. In another seven the primary reason was difficulties with adherence.

Twelve out of the 40 in fact declined the offer of injectables and one moved out of the area. One is yet to start, while one was already receiving injectables because they were in the FLAIR study. This meant that 13 men and 13 women have so far started receiving injections, with ages ranging from under 24 to over 65. The average duration so far on injections is 13 months.

Two people experienced virological blips but without needing to change treatment. One person had a viral load of 70 at their third injection and 176 at their sixth. They have since remained suppressed for four more injections. Another had a viral load of 359 at their second injection, but it was decided to continue. This became undetectable again for the next four injections (nine months) but at the time of presentation, had just turned up for their fifth with a viral load of 200, so it is not yet known if they will re-suppress.

So far, only two people have had to stop. One was a woman who became pregnant. The other, however, had virological failure with resistance to both drugs. He had had a viral load of just over 50 at his first injection but this was seen as a blip. Tested on the day of his fifth injection, he turned out to have a viral load of 953 at the time, which on re-testing two weeks later was 1050 and then a week after that was 1800. At this point he was put on Symtuza, the same oral regimen as one of the breakthrough patients described above.

This person turned out to have HIV with a resistance mutation to rilpivirine called E138K, which is the classic one encountered in cases of resistance to oral rilpivirine, but also had three resistance mutations to integrase inhibitors at positions 140, 148 and 155. He had HIV which was a mixture of wild-type and resistant virus, generally a sign that it had developed quite recently. Because he had been virally suppressed since 2009, resistance tests hadn’t been possible for several years, but in 2009 he had definitely not had the rilpivirine resistance. Although he was not tested for integrase resistance then, the drug class had only just been introduced so it is most unlikely that had acquired transmitted resistance.

Dr Drury described the effect of the failed injectable therapy as devastating for the patient. “He was someone who experienced having to take daily ART as depressing and was looking forward to being freed from this burden,” she said. She added that a friend of his, who had also been considering injectable ART, had decided not to take it when his therapy failed.

More understanding of injectable failure needed

It is important to emphasise that breakthrough infections on injectable therapy are still rare, with only one per 80-100 people starting. However, there is still much that is left to understand about why injectable therapies occasionally fail in conditions of perfect adherence. Neither of the two patients in this report had HIV of the A6 subtype that seems to increase the risk of failure or had any of the other predispositions for failure cited in this report, though we don’t have detailed data of their drug levels. It does seem to be the case that viral blips, even to quite low levels, have an association with subsequent failure – especially ones found at the crucial time of the first injection. But to take all ‘blippers’ off injectables would rule out 5% of people taking them.

There is also quite a lot we do not understand about resistance to integrase inhibitors, including cabotegravir. At last year’s CROI, one study found unexpectedly that some mutations in HIV’s env protein, rather than in its integrase enzyme, may lead to integrase resistance as they facilitate a form of ‘escape’ from the effect of the drug.

A poster at BHIVA 2024 found that there was a higher rate of failure in patients starting ART (usually, oral therapy) who had one particular env mutation called A539V. This is uncommon (35 occurrences in 814 people) but is more common in people of African origin or who have a subtype of HIV called CRF02_AG. The rate of viral non-suppression in people on ART 12 months after starting was 29% in people who had A539V, compared with 12% in people who did not have it. Although not all these people were taking integrase inhibitors, A539V plus any mutation in the integrase enzyme greatly increased the likelihood of viral failure.

It appears that doctors may have to spread the net wider in testing for warning signs of the possibility of failure in these otherwise welcome and well-tolerated regimens.

Correction: This article was amended on 9 May 2024. An earlier version named the combination of cabotegravir and rilpivirine as Apretude, rather than Cabenuva.