Two-monthly HIV injections matched oral therapy in people with adherence problems

Design

Phase 3b, randomised, open-label non-inferiority trial, 48 weeks

Population

540 virally suppressed adults with HIV-1 and adherence challenges, Kenya, South Africa, Uganda

Primary outcome

Viral load <50 copies/mL at week 48

Effect

91.0% vs 89.2% (difference 1.9, 95% CI -3.1 to 6.9); virological failure 1.9% vs 0%

IMPALA, a phase 3b open-label trial at seven sites in Kenya, South Africa and Uganda, randomised 540 adults with HIV-1 who were virally suppressed but had a history of suboptimal adherence or care engagement. They switched to intramuscular cabotegravir plus rilpivirine every eight weeks or continued daily oral dolutegravir-based therapy. Baseline resistance testing was not done. Published in the New England Journal of Medicine on 15 October 2026.

At 48 weeks, viral load was below 50 copies in 91.0% on injections and 89.2% on oral therapy (difference 1.9 points, 95% CI -3.1 to 6.9), meeting non-inferiority. Confirmed virological failure occurred in five people on injections (1.9%) and none on oral therapy. Serious adverse events were similar.

This matters because the trial deliberately enrolled the people long-acting therapy is meant for, in a setting without routine resistance testing. The failures are a reminder that rilpivirine resistance and missed injection windows can undo the regimen.

In India, long-acting cabotegravir-rilpivirine is not part of the national programme and access is limited; its regulatory status here should be checked before discussing it with patients.

• Consider long-acting injectable therapy for suppressed patients who struggle with daily pills, where available
• Check for prior non-nucleoside resistance or treatment failure before switching, since rilpivirine is part of the regimen
• Build a recall system so injection visits are not missed
• Monitor viral load closely in the first year after switching
• Keep dolutegravir-based oral therapy as the standard where injections are unavailable

Why it matters

Long-acting therapy worked in the population and setting it was designed for, not just in highly adherent trial volunteers.

Don't overread it

Five virological failures against none on oral therapy is a small number but a real signal; resistance data are needed.

The statistics, in plain English

The lower bound of the confidence interval (-3.1 points) stayed well above the -10 point non-inferiority margin, so injections were not worse on suppression. The virological failure difference has an interval of 0.3 to 3.5 points, so it is unlikely to be chance.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.