Prevention is a huge theme of this AIDS conference, as it was in Toronto. Today, scientists addressed one of the uncertainties that might sandbag a promising prevention strategy.

Many scientists, advocates, public health officials hope that antiretroviral drugs, given as topical microbicide gels or as oral “Pre-Exposure Prophylaxis” (PrEP), might protect people from HIV infection. Twelve microbicides are being tested in clinical trials, and seven trials of oral pills as preventatives (a.k.a. pre-exposure prophylaxis, or PrEP) are planned or underway. Despite some early failures, many HIV experts are optimistic that ongoing trials will yield more positive results.

But one big issue looming over these trials is drug resistance. The drug regimens used in the trials aren’t designed to stop HIV from multiplying in the body – they’re only designed to keep the virus from invading in the first place. So there’s concern that people infected with HIV will develop drug resistance if they’re inadvertently placed on preventive drug regimens. That could be a problem for both HIV-infected patients and anyone else to whom they transmit a resistant virus.

Today, John Mellors of the University of Pittsburgh showed a mathematical model that describes how such resistance might evolve, and how serious a problem it could be.

Mellors’s model predicts that the amount of drug resistance in these prevention trials will depend on lots of factors, including how many people with HIV unknowingly participate in the trials. In the best case scenario, relatively few will enroll, and no serious drug resistance will arise, he said. But if many HIV-infected people end up taking the drugs – the worst case scenario – its potential benefits to a population will be nullified by the amount of drug-resistant virus unleashed on that population.

Mellors said his study yielded one take-home message: “PrEP rollout must include routine testing” – as often as once a month – to ensure the people taking it are still HIV-negative. “If we give PrEP indiscriminately to those infected, we will have much higher resistance,” Mellors declared.

And, as Mellors pointed out, regular HIV testing might be possible in the context of a clinical trial, in which study volunteers agree to be strictly monitored. But it’ll be a lot more difficult in real life, when many more people take the drugs under less controlled conditions.

The scientists running these prevention trials already know that drug resistance might be a concern, and they’re watching out for it. But Mellors’s study suggests we should already be thinking about what to do next if these trials succeed. We’ll need a plan for giving out the drugs responsibly. And given how difficult it can be to convince people to be tested for HIV even once, testing them over and over again could be challenging indeed.

It’s just one more reminder of a mantra that has been repeated often at this conference: we can’t count on any magic bullet to stop HIV.