Michael Gwarisa
For years, the global HIV response has rallied around a simple but powerful message: Undetectable equals Untransmittable, widely known as U=U. Backed by strong scientific evidence, the concept has transformed both HIV prevention and the lives of people living with the virus. But new findings from the Rakai HIV Cohort Study in Uganda suggest that the story becomes more complex when viral load is low but still detectable.
Researchers analysing data from the long-running cohort found that while most HIV transmissions still occur at higher viral loads, a small proportion, about 8 percent, were linked to individuals whose viral load ranged between 200 and 1000 copies per millilitre. These levels are considered low, but not undetectable, raising important questions about how risk is communicated just above the U=U threshold.
The findings were presented at the Conference on Retroviruses and Opportunistic Infections 2026 by Michael Martin of Johns Hopkins University. Using modelling based on observed transmission patterns, the study estimated that individuals with low but detectable viral loads still carried a measurable risk of passing on HIV over time.
According to the model, a person with a viral load of 15,000 or more had about a 14 percent chance of transmitting HIV to a partner within a year. That risk dropped significantly to 2.9 percent at a viral load of 1000. At 200, the lower boundary of what is considered undetectable in U=U messaging, the estimated annual risk was 0.6 percent.
The Rakai cohort itself is one of the most influential HIV studies in the world, involving around 15,000 heterosexual couples in Uganda where one partner is HIV positive and the other is not. Earlier research from the same cohort, published in 2000, found no transmissions when the HIV-positive partner had a viral load below 1500 copies. That finding helped shape early understanding of how viral load affects infectiousness.
Since then, major international studies such as PARTNER 1, PARTNER 2 and Opposites Attract have strengthened the scientific foundation of U=U. Across more than 150,000 instances of condomless sex among both heterosexual and same-sex couples, researchers found zero cases of HIV transmission when the HIV-positive partner maintained a viral load below 200. Crucially, genetic analysis confirmed that no linked transmissions occurred in those studies.
The new Rakai analysis differs in important ways. Unlike the earlier U=U studies, it did not include phylogenetic testing to confirm whether transmissions occurred within the primary couple. Instead, it relied on participants reporting that they had no other sexual partners. That limitation means some infections could have originated outside the relationship.
Researchers behind the study have been careful to stress that their findings do not challenge U=U itself. Alison Hill, who led the analysis, said the goal was not to question the principle that people with undetectable viral loads do not transmit HIV. Rather, the study sought to understand what happens in the less clearly defined range between full suppression and higher viral levels.
Their focus was on what scientists call low level viraemia, where the virus is present in small but measurable quantities. This range, between 200 and 1000 copies, can be unstable. Viral load may be rising due to treatment interruption or adherence challenges, or it may be falling as treatment takes effect.
This instability is central to interpreting the findings. In several of the observed transmission cases, viral load measurements were taken months before the HIV-negative partner tested positive. On average, there was a gap of about 141 days between tests. That is enough time for viral load to change significantly, meaning it may have been higher at the actual moment of transmission.
Similar patterns have been observed in earlier trials. In one case from the HPTN 052 study, a transmission occurred from a partner whose last recorded viral load was relatively low. However, further analysis suggested the individual likely had a higher viral load at the time of transmission due to a rebound.
The Rakai researchers also explored this uncertainty by modelling transmission risk among individuals not on antiretroviral therapy. In this group, viral load tends to be more stable than in those on treatment, where interruptions can lead to fluctuations. Even in this more controlled scenario, the estimated transmission rate for low level viraemia was about 1.3 per 100 person-years.
The findings come amid ongoing discussions within the World Health Organization about how best to define and communicate viral suppression, especially in lower-income settings where some tests do not detect viral loads below 1000 copies. A recent WHO review identified a small number of cases where transmission appeared to occur within this suppressed but detectable range.
For clinicians and public health experts, the message is not to undermine U=U, but to sharpen its boundaries. U=U applies specifically to people who are on effective treatment and have a consistently undetectable viral load below 200 copies. Within that range, the evidence remains clear and robust: sexual transmission does not occur.
What the Rakai study adds is a clearer picture of the grey zone just above that threshold. It shows that while risk drops dramatically as viral load decreases, it does not fall to zero until the virus is fully suppressed.
For people living with HIV, the implication is straightforward but important. Staying on treatment and maintaining an undetectable viral load remains the most effective way to protect both personal health and partners. For policymakers, the findings highlight the need for accurate testing, consistent treatment access and careful messaging that reflects both the power and the limits of current science.
In the end, the science still supports U=U. But as this study shows, understanding what happens just above “undetectable” is essential to closing the remaining gaps in HIV prevention.
ORIGINAL ARTICLE IN THE AIDSMAP
