How We Know HIV PrEP Works and Why Adherence Changes the Number

A percentage on a fact sheet always hides a condition. When the Centers for Disease Control and Prevention says PrEP (medicine taken before possible contact with HIV) lowers the risk of getting HIV from sex by about 99 percent when taken as prescribed, the load-bearing phrase is the last one. That figure is not a permanent trait of a molecule. It is what the molecule delivers when it is actually present in the bloodstream at the moment it is needed. The whole scientific story of PrEP is the story of researchers learning to tell two questions apart: does the drug work, and does the person take it.

Key points#

• The roughly 99 percent figure applies to people for whom the drug is reliably in the body; it is a ceiling, not a guarantee.
• Early oral PrEP trials looked weak on average because many participants had no detectable drug in their blood, not because the medicine failed.
• Once daily PrEP was proven, a placebo group became unethical, so newer trials compared against an active pill and an estimated background infection rate.
• Long-acting cabotegravir and twice-yearly lenacapavir work largely by guaranteeing drug levels that daily pills often miss.
• The CDC graded the twice-yearly lenacapavir evidence as high certainty, language that reflects trial design, not marketing.

What the number is really measuring#

Think of PrEP protection as the product of two things multiplied together: how good the drug is, and how consistently it reaches a protective level. A superb drug taken erratically and a weaker drug taken perfectly can land at the same real-world number for very different reasons. This is why a single reported percentage can be both true and incomplete. It answers the question "what is possible" without answering the one you probably came for, "what happened in this population."

The distinction has a formal name. Efficacy is what a treatment can achieve under ideal, measured conditions. Effectiveness is what it achieves in ordinary use, where doses get skipped, refills lapse, and life intervenes. For most medicines the gap between the two is modest. For daily oral PrEP it turned out to be enormous, and closing that gap has driven nearly every advance in the field.

The early trials hid their best result inside an average#

The first large randomized, placebo-controlled trials of tenofovir-based oral PrEP were almost misleading at first read. The iPrEx trial, published in 2010, enrolled men who have sex with men and transgender women and reported that daily emtricitabine-tenofovir cut HIV acquisition by only about 44 percent against placebo in the intention-to-treat analysis. Taken alone, that number would make PrEP sound like a coin flip.

The average concealed the real finding. When investigators measured drug in participants' blood, roughly half had none detectable at the moments they were sampled. Among the participants who did have detectable drug, the estimated risk reduction was about 92 percent. Pharmacology substudies tied intracellular tenofovir concentrations consistent with four or more doses a week to strong protection. The compound was potent. The bottleneck was the daily act of swallowing a pill.

The same signature appeared across the whole research program. In serodifferent heterosexual couples, the Partners PrEP study reported high efficacy where measured adherence was high. Studies in populations that struggled with daily dosing produced low or even null results, driven not by a failed drug but by low drug levels. In people who inject drugs, the Bangkok Tenofovir Study anchors the CDC's at-least-74-percent figure for that route, and there too protection climbed with consistent use. The recurring lesson is blunt: the PrEP percentage you read is a joint statement about a drug and a behavior, never about the drug alone.

When you cannot ethically use a placebo anymore#

Proving that daily oral PrEP works created a new problem for anyone testing the next drug. Once an effective prevention tool exists, assigning volunteers to a sugar pill and letting some acquire a lifelong infection is no longer acceptable. Newer trials therefore had to measure a prevention drug without a placebo arm.

They solved this with two reference points. The first is an active control, meaning an already approved oral regimen that the new drug must beat or match. The second is an estimated background incidence, the rate of new infections expected in a comparable untreated group, often reconstructed with recency assays that separate recent infections from long-standing ones. That combination is what let the long-acting studies claim genuine superiority rather than mere non-inferiority.

Removing the daily decision from the equation#

If skipped pills were the true limiting factor, then a drug that removed daily dosing should reveal the compound's full strength. That is exactly what happened. Long-acting cabotegravir, injected every two months, was tested head to head against the daily oral pill. In HPTN 083, among men who have sex with men and transgender women, the injection produced roughly one-third the infections seen in the oral group. Its companion trial, HPTN 084, showed a similarly large advantage in cisgender women.

The injection did not smuggle in a categorically stronger molecule. It guaranteed the drug levels that a daily pill so often failed to sustain in real life. Most of the improvement was the difference between what the pill could do and what it actually did once handed to a busy human being.

Twice-yearly lenacapavir and the phrase "high certainty"#

The logic then scaled to an injection given under the skin just twice a year. Lenacapavir, a capsid inhibitor, was studied in the PURPOSE program. In PURPOSE 1, among adolescent girls and young women in South Africa and Uganda, there were zero HIV infections among the 2,134 participants who received it, an incidence well below both the reconstructed background rate and the oral comparator. In PURPOSE 2, among men and gender-diverse people, two infections occurred among 2,179 participants, so about 99.9 percent stayed uninfected, roughly a 96 percent reduction against background incidence.

The Food and Drug Administration approved twice-yearly lenacapavir in June 2025 as the first PrEP option dosed every six months, and in September 2025 the CDC recommended it, citing a high certainty of evidence for its efficacy and safety. That grading is not a flourish. Under structured frameworks such as GRADE, certainty rises when the design is a randomized trial with a hard, objective endpoint (documented HIV seroconversion), when enrollment is large, when the direction of effect is consistent across studies, and when the estimates are precise. The PURPOSE program met those conditions, which is how a recommendation earns high certainty rather than moderate or low.

Reading the figure honestly#

For a daily pill, being reliably present depends on a choice you make every morning. For a bimonthly or twice-yearly injection, the same reliability becomes a scheduling and access question settled at a clinic visit rather than re-litigated every morning. That is the underlying reason the newer options narrow the distance between trial performance and everyday performance. The near-99 percent belongs to the person for whom the drug is genuinely onboard, and the value of long-acting formulations is that they make that condition far easier for you to meet.