Possible, but impossible

One of the major takeaways from the International Conference on HIV Treatment and Prevention Adherence, where I am at, was pretty much laid out in the beginning by keynote speaker Wafaa el-Sadr. She is among other accomplishments the winner of one of those McCarthur Foundation genius awards, so I guess she must be a genius. So listen up.

We were discussing a few days ago the eradicability of certain diseases. Smallpox has already been eradicated, and we've been just on the verge with polio for a few years now. (It's very disappointing that we haven't made it, due mostly to political obstacles, and it will be a catastrophe, in my view, if polio again becomes widespread and we have to once again start immunizing kids all over the world. So let's get this thing done.) Jimmy Carter, who is an excellent former president, is behind the near eradication of the guinea worm, which will likely be completed soon enough.

To be eradicable, a disease needs to have no non-human reservoirs. Guinea worm spends part of its life cycle in water, but needs a human host to complete it. There also must be a means either to prevent transmission. A vaccine is most convenient, but in the case of the guinea worm it's done by purely mechanical means: filtering drinking water, and making sure that the worms, when they emerge from their host, do not have an opportunity to get back in the water and reproduce.

There is no vaccine for HIV, but it turns out -- and people really weren't sure of this until now -- that if an infected person rigorously adheres to an anti-retroviral medication regimen, so that there is no detectable virus in the blood plasma, there will also be no virus in the genital mucosa and secretions and they will be almost entirely non-infectious. We know this from studies with sero-discordant couples. Evidence for non-infectiousness by sharing injection equipment is only circumstantial and there are good reasons to think it may not be so complete. Also injection drugs users are less likely to be adherent to their medications.

Nonetheless you don't need to achieve 100% non-transmission to stop an epidemic. If transmissibility is low enough, chains of infection will be short and the pathogen will eventually die out. Also, if people are in treatment it's much easier to augment prevention by education and needle exchange. People who are in treatment are actually less likely to engage in transmission risk behavior. This may seem counterintuitive, but basically, it means they have come to terms with their situation and are dealing with it. They don't want to be reinfected or get an STI, and they don't want to infect others. If they are addicts, they are more likely to stop or at least practice harm reduction.

But, can we treat our way out of this epidemic? Almost certainly not. Here's why.

First, everybody who is infected will have to know it. In the U.S., 20% or more of infected people are unaware of their status. The number is higher elsewhere, particularly in sub-Saharan Africa and South Asia. Then, every one of them must be immediately linked with medical care. You get a big fall of there, even in the wealthy countries where many people go through an initial period of avoidance and denial.

Then, they have to get treatment. This is tricky because it isn't clear that it's in the individual's own interest to begin treatment immediately. Right now the standard of care calls for initiation when the CD4+ cell count -- those are the kind of white blood cells that are preferentially infected and destroyed by HIV -- falls below 350/ml. There are indications that long term outcomes might be better if you start earlier, around 500, but it's not clear that it makes sense to start earlier. The pills have long-term side effects that can eventually catch up to you, mostly abnormalities of lipid metabolism that can cause disfiguring fat redistribution, weight gain, hypercholerstolemia, and diabetes. Also there is the danger of acquiring drug resistant virus if adherence is less than perfect. It isn't considered ethical to ask people to do something for the benefit of public health that might harm them personally, and a lot of people probably wouldn't do it anyway. You can intervene to influence behavior at that point, but you'll never be 100% effective.

Then, the people need to adhere strictly to their prescribed regimen, which is difficult to do and a lot of people don't do it. Recent data indicate that if you miss two or three days consecutively on typical regimens, you risk viral rebound. Longer treatment interruptions carry even higher risk.

So, right now, in the United States, only about 19% of HIV infected people have suppressed viral load and are essentially non-infectious. In Africa, of course, it's far worse, and in fact we're falling further behind as the rate of new infections exceeds the rate at which people can be offered treatment. And treating people earlier means few people could be treated altogether given fixed resources.

So no, we won't eradicate HIV by treating it. But, the more people we can treat, the fewer new infections there will be. It's a virtuous circle, even if it won't be entirely closed. That makes it a good investment. But like many good investments, the depraved state of our current politics means we aren't going to make it.