Broadly Neutralizing Antibodies and Viral Rebound: Early Signals from the RIO Trial at CROI 2026

At the 2026 Conference on Retroviruses and Opportunistic Infections (CROI), investigators presented preliminary data from the RIO trial suggesting that exposure to broadly neutralizing antibodies (bNAbs) may alter viral rebound dynamics following treatment interruption in a subset of individuals living with HIV.

The findings do not represent a cure, nor do they support treatment interruption outside clinical trials. However, they provide early evidence that immune-based interventions might modulate post-treatment viral control.

The results presented reflect only part of the ongoing study, and several mechanistic analyses remain in progress.

Study Design and Rationale

The RIO trial evaluates two long-acting broadly neutralizing antibodies — 3BNC117-LS and 10-1074-LS — administered to individuals who initiated antiretroviral therapy (ART) early after infection and maintained sustained viral suppression.

Broadly neutralizing antibodies bind conserved regions of the HIV envelope, blocking viral entry and potentially enhancing immune-mediated clearance of infected cells. The LS modification extends half-life, allowing prolonged circulation.

The study includes analytically supervised treatment interruptions (ATI), conducted under strict safety criteria.

Participants were randomized 1:1:

• Arm A received dual bNAbs.
• Arm B received placebo.

In the first ATI phase (ATI-1), viral rebound occurred as expected across participants. In Arm B (placebo), rebound was rapid and universal. By week 10 of ATI-1, all participants had met rebound criteria.

Viral rebound was defined as either >1,000 copies/mL for six consecutive weeks or >100,000 copies/mL for two consecutive weeks, with adjudication in special cases.

This confirmed the established observation that ART cessation results in detectable viremia within weeks.

Delayed Viral Rebound in ATI-2

Participants in one arm of the study (Arm B) initially received placebo during ATI-1. After ART re-suppression, they were administered the dual bNAb combination. Following an antibody washout period of approximately 24 weeks on ART, a second treatment interruption (ATI-2) was performed.

This within-participant comparison revealed a marked shift in rebound kinetics.

During ATI-1, 100% of participants experienced viral rebound by week 20. During ATI-2, approximately 50% had not rebounded by week 20. Among those with delayed rebound, viral load rose more slowly and peaked at lower levels.

Notably, delayed rebound occurred despite only modest circulating antibody concentrations at ATI-2 start (median just under 50 µg/mL), similar to levels observed at rebound in Arm A during ATI-1. This suggests the effect may not be explained solely by passive neutralization.

Of 34 Arm B participants, 28 were included in the paired ATI-1 versus ATI-2 analysis. Two individuals with ongoing viral control during ATI-1 for nearly four years — likely post-treatment controllers — were excluded from this comparison.

As Julia Edgar noted during the discussion, “We are actively working to understand what factors distinguish those who control from those who do not — it’s unlikely to be a single mechanism.”

Is the Effect Attributable to Antibodies?

An important question raised during the session concerned causality.

Jonathan Lee (Brigham and Women’s Hospital) asked whether repeated treatment interruptions alone might alter rebound kinetics, referencing earlier ACTG data showing that sequential ATIs can influence viral dynamics.

Edgar acknowledged the analytical complexity: “It’s a bit tricky of a comparison because of the washout period and differences in antibody concentration at the start of ATI-2. But we are looking at it.”

Distinguishing between an effect of repeated ATI and a true antibody-mediated modulation remains central to interpreting the data.

If confirmed, the findings would support the hypothesis that bNAbs can induce durable immunologic changes beyond transient viral neutralization.

Mechanistic Questions: Resistance, Sensitivity, Immune Modulation

Preliminary analyses from Arm A, conducted by the Rockefeller University group, identified resistance mutations emerging primarily against 10-1074 following rebound.

Whether delayed rebounders exhibit distinct resistance patterns, altered inflammatory markers, enhanced T-cell or NK-cell responses, or reservoir changes remains under investigation.

Investigators defined a 9-week threshold for delayed rebound based on the mean time to rebound plus two standard deviations during ATI-1. Using this data-derived cutoff, two rebound phenotypes emerged in ATI-2: early rebounders (<9 weeks) and delayed rebounders (>9 weeks).

Another hypothesis raised during the discussion was viral hypersensitivity: could delayed rebound reflect unusually high sensitivity of residual virus to modest antibody concentrations?

“We don’t yet know the sensitivity profile of the viral reservoir before ATI-2,” Edgar responded. “There is likely significant inter-individual variability.”

Current analyses are evaluating viral genetics, immune phenotyping, and functional responses to determine whether the observed effect reflects persistent low-level neutralization, altered reservoir composition, immune priming, or a combination of mechanisms.

At present, no definitive mechanism has been established.

Ethical Considerations in ATI Design

The trial also prompted discussion regarding the ethics of placebo-controlled ATIs.

Community representative Peter Allen questioned whether additional placebo interruptions are justified, given that viral rebound in untreated interruption is already well characterized.

Investigators noted that when RIO was designed, controlled placebo data were still limited. As the evidence base grows, the field may reconsider whether new placebo arms remain necessary.

The exchange underscored a broader shift in HIV cure research: balancing mechanistic rigor with participant risk minimization.

Interpretation and Implications

The RIO data suggest that dual broadly neutralizing antibody exposure may modify viral rebound kinetics in approximately half of treated individuals under controlled conditions.

This does not equate to durable remission. ART remains the standard of care.

However, the findings raise the possibility that antibody-based interventions may exert immunomodulatory effects capable of shifting host–virus equilibrium, at least transiently.

The magnitude and durability of this shift — and whether it can be enhanced through combination strategies — remain open questions.

As presented, these results represent an early signal, not a conclusion.

But in a field where functional cure remains elusive, even modest changes in rebound kinetics warrant close examination.

Further data from additional study arms, including Arm C designed to assess the contribution of ATI alone (“autovaccination”), and ongoing mechanistic analyses will be critical to determine whether RIO marks a meaningful advance in immune-based HIV control.