Taking anti-HIV pills every day cuts the risk of infection by 63%, said CDC researchers in Botswana. It cuts infection by up to 73%, said University of Washington researchers working in Kenya and Uganda. That’s great news, of course.

Here’s why I’m confused. The larger of these trials was conducted in 4,758 “discordant couples”. [I earlier incorrectly reported that both trials were in discordant couples. The CDC trial in fact recruited 1,200 sexually active uninfected heterosexuals, regardless of their partner status. Full inclusion and exclusion criteria here]. That means researchers in the large discordant couple trial knew that one person was infected and the other uninfected. They chose to give drugs to the uninfected person, to see if it would stop them becoming infected. And it does, in over 60% of cases. But another recent study shows that if we give the drugs to the infected partner, the one who might actually need these same drugs because they have HIV and need it surpressed, it cuts infection by 96%. So in the case of discordant couples, it seems to make much more sense to give the antiretrovirals in question to the infected partner.

That leaves us with the question: who should get PReP? Right now, there are not enough antiretrovirals to go around to treat all the sick people who need treatment. If we’re going to use them selectively for prevention, we should start with the most effective use, which appears to be early treatment of the infected partner in discordant couples. We could also give them to people who aren’t in a couple but who know that they’re likely to get around a bit and might want to stay safe without using condoms. That’s potentially a lot of people; it will stretch our purses. But more than that, it will stretch our political will. Let’s face it, HIV has reached eye-watering levels in many sub-Saharan African countries because both voters and governments have been in deep denial about their own, and their neighbours’, propensity to have sex with someone who is not their single life-time partner. Some people, including influential religious and community leaders, even continue to believe that giving out condoms encourages licentious sex. To them, giving out ARVs will surely mean encouraging licentious unprotected sex (if you’re anti-condom, is that better or worse?).

So who is PReP for? We’ve got a better option for discordant couples. We’re not going to want to give it to randy adolescents. We know it works for gay men, but some of the countries where the trials took place would rather thump or jail gay men than protect their sexual health. We’ve no idea yet if it works for drug users (though a deeply unethical trial by CDC in Thailand will tell us that soon.

Of course PReP will find its niche; when people actually take it it works really well (though not as well as abstinence, when people actually abstain, or condoms, when people actually use condoms). We’ll find out a bit more about just how well at the annual AIDS circus in Rome next week. I’ll look forward to learning what the actual incidence rates in the studies were, and more about sex differentials and adherence. But I think we would be unwise to rush around talking about massive roll-out of PReP before we actually figure out who it works for in the real world.

As an aside, the results have a huge potential impact for Gilead, manufacturer of both Viread (bascially tenofovir, one of the pills that worked in the trial) and Truvada (the tenofovir – emtricitabine combination that was the other). Gilead has come over all generous and has started letting Indian and other developing country companies copy their products. They’ll take a 5% fee; if we really do go for a massive roll-out of PrEP, that will keep drug costs down globally, while giving Gilead extra cash for very little effort. A win-win situation for which they should be congratulated.

A second aside: The CDC trial is confusing in a different way. In December 2009, CDC announced it was terminating the trial of Tenofovir for HIV prevention because they’d had so many drop-outs that the trial would be unlikely to show results even if they doubled the size of it. They kept it going not as an efficacy trial (testing Tenofovir against a placebo) but as a safety and behavioural trial (clocking how good people were at taking their pills, looking for side effects etc.). So it was quite surprising to find them leaping forward with efficacy reults, of which more details here.

Thanks to Eva for pointing out my error.

That’s huge. So huge that the study was stopped early. We still don’t have many details about things that I would find interesting — how good were people at taking their pills, did they people on meds have more or less unprotected sex than people who weren’t on pills, etc. — but it seems incontrovertible that if you’re infected with HIV, one way to protect your sex partners is to start taking antiretrovirals when you’re immune system is still in relatively good shape.

I’m still left with two major questions. First: it is clear you’ll protect your partners, but will you protect yourself? What do we really know about the long-term effects of taking antiretorvirals early for your partners’ benefit? We’ll get more information about that from another trial by the same group, but they’re not scheduled to report for another five years. People who got treated earlier in the the treatment-as-prevention trial were just as likely to die during the course of the study as those who didn’t, though encouragingly, they were signigicantly less likely to get sick with TB. It may well be that starting meds earlier is good for the infected person as well as for those they shag.

Second major question: this study (known as HPTN 052) has made it clear that an HIV infected person whose CD4 count is between 350 and 550 when they start treatment is less infectious than a person who doesn’t start until their cell count falls below 250. For those individuals, treatment is prevention. But does that necessarily mean that expanding treatment will reduce new infections at a population level? For an interim period, at least, it may well not. Before I’m accused of raining on the parade yet again, I want to point out that the same question was raised by Myron Cohen, the principle investigator of HPTN 052 in an e-mail exactly a year ago. Speaking of a stampede towards using earlier treatment as a means of prevention, in part as a result of a lot of “utopian” modelling, Myron said:

“I am not convinced that this will all come out the way it now appears, and we do not yet know how to measure population level benefit of ART, if it is to occur.”

Logically, if you reduce the infectiousness of every infected person by 96%, new infections will fall very dramatically. But we know that can’t happen. It certainly can’t happen overnight. It’s worth noting that genetic analysis of the virus shows at least 18% of the new infections in the study (and possibly up to 28% — not all the analysis is finished) came from someone who was not the “regular partner” recruited into the study. Until everyone gets treated sooner, those infections will continue. Indeed some will continue even with universal earlier treatment, because some will probably have come from people who are newly-infected, very infectious and unlikely to be treated. That’s a continuing worry in places where the all-too-visible face of AIDS-related emaciation, disfigurement and death prompted a change in behaviour; less sex, fewer partners, more condoms. As expanded treatment removes that visible death-mask, communities revert towards pre-AIDS behaviours. Where condom use rose rapidly, for example among gay men in rich countries, it has fallen back since relatively early HIV treatment has been universally available. The effect may be less pronounced in the hyperendemic countries where behaviour has not changed all that much, but it’s something to watch out for. More unprotected sex with a variety of partners also pushes up STIs, and an active STI can in turn unleash spikes of HIV in the genital fluids and undermine the protective effect of antiretrovirals. Note that I’m not talking here about the behaviours of discordant couples who have been to counselling and are on HIV treatment, I’m talking about people who believe (or assume, or just hope) that both they and their partners are negative.

It’s perhaps worth clocking that researchers shifted their original “deferred treatment” threshold to a CD4 count of 250 (from 200) when the WHO treatment guidelines (and the national guidelines of many countries they were working in) shifted. They did not, however, change it again when WHO guidelines were revised upwards again to 350, because “the second revision was not readily adopted by all of the countries participating in the study, primarily due to a lack of drug supply.”

The fact that 1.8 million people died of AIDS in 2010 confirms that many countries have trouble getting drugs even to those people who depend on them for survival. Getting them to the larger number who might benefit from them as a transmission risk and TB reduction measure will be harder still. That will eat into the potential prevention gains in two ways — obviously people who don’t have drugs don’t have lower viral loads. But relatively healthy people who sometimes have drugs may present more of a transmission risk than those who never do, because HIV tends to spike upwards into a brief, highly infectious phase when treatment is interrupted. Frequent interruptions can undermine the effectiveness of the drugs; resistance is another source of nasty, infectious spikes in viral load. Though we don’t yet have any information about adherence, we can assume that people in the HPTN 052 trial had uninterrupted access to meds, and we know from the study protocols that they were actively encouraged to keep taking them. We also know that they deliberately excluded drunks, people with drug problems, people with mental problems or “Any condition that, in the opinion of the study staff, would make participation in the study unsafe, complicate interpretation of study outcome data, or otherwise interfere with achieving the study objectives”. In the real world, we can expect a more erratic drug supply, sloppier adherance and bouncier viral loads. That may will turn in to protection of far less than 96%.

Overall, more people on treatment means, we hope, more people living longer, healthier, more sexually active lives. That also means more opportunities for sex with someone when viral load is spiky, and thus for onward transmission over HIV. Add together more unprotected sex with people who may be in not-yet-treated primary infection, and more sex during times after the start of treatment when HIV is bouncing around because of STIs, treatment interruption, treatment failure or whatever. If the sum of those two adds up to more than the sex a person has between the time their CD4 count hits 550 and the time it would otherwise have hit 250, new infections are likely to rise, even if earlier treatment reduces transmission during that notional window to zero.

That is absolutely no reason at all not to push to use antiretrovirals to reduce infectiousness in people who are infected. I am persuaded that we should be doing that, and I think HPTN 052, with its relatively sober threshold for starting even the “early” treatment points us in the right direction. But I think it will be a long time before we have the cash and the systems in place to make this an effective prevention tool at the population level. And since none of the other prevention tools we have are working very well at the population level either (at least in unpaid sex of any persuasion), we certainly can’t declare victory quite yet.

So gay guys go on having unprotected sex after they are diagnosed with HIV, a new descriptive study of gay poz guys at a clinic in Boston tells us. Nothing new there, although it’s sobering to be reminded that one in two of the men who know they have HIV choose to bareback with someone who may be negative.

The most important finding from the Boston study is that the more recently diagnosed a guy is, the more likely he is to be exposing other people. We can’t tell from the paper if that’s something new. It may be that there’s a blast of screwing around soon after diagnosis, possibly as a reaction to it, then a calming down. The post-diagnosis binge is one of the possible explanations given by NAM’s ever-sensible Gus Cairns, posting over at rectal microbicide site IRMA.

The thing that most determines whether a poz guy will pass on HIV during unprotected sex is his viral load. We know that’s likely to be highest for the few months after he first gets infected. So if we do get better at detecting HIV early and we don’t do anything about bringing viraemia down instantly, the “post-diagnosis sex binge”, if real, would be a worry we’d have to add to the known dangers of the “slutty phase” that affects undiagnosed men as well.

Gus gives us another potential explanation:

3. The historical explanation. It can’t be about young gay men not having experienced AIDS, because it was recent diagnosis that was the risk factor, not age. But it could be about prevention fatigue that affects all ages, and lack of relevant, effective and up to date messages. The result would be that the recently diagnosed have higher risk behaviours (and have caught HIV) because they haven’t internalised prevention messages in the way that the longer-term diagnosed seem to. Does risk behaviour decline over time in the longer-term diagnosed for one reason or another (more ease of disclosure, self-education, awareness of criminalisation, catching one too many STIs, etc) or will men diagnosed today continue to be higher-risk than men diagnosed years ago?

If it’s a binge thing, behaviour will get safer over time. If it’s an “I’ve zoned out HIV messages” thing, it won’t. But there might be something else going on: The “HIV isn’t so painless after all” thing.

One of the reasons that HIV prevention messages are failing is that much of the public health world still treats HIV as though it’s AIDS, as though it is self-evident why you would want to prevent it. But now that AIDS has virtually disappeared, what’s the big deal about HIV? Why bother to protect yourself, or to avoid passing it on?

Guys who were diagnosed longer ago are more likely to have realised that HIV (like diabetes and arthritis) is actually more than a one-pill-a-day shrug-off. The ups and downs of treatment — having to call off a date because you’ve blown up like a tomato, worried that your boss will see you popping pills, having to cancel a day’s skiing because you’ve got to go for your viral load monitoring, — it can be a real pain. A pain that, on reflection, you might go out of your way to avoid passing on. That may be one reason why people who were diagnosed longer ago are less likely to expose their partners to HIV.

But treatment is improving all the time; as prevalence goes on rising and the ick factor falls, HIV becomes less and less of a pain. It seems likely to me, then, that barebacking will continue to rise. That makes people in public health crazy, of course. We have to think about resistance, a reappearance of AIDS, costs to the health system. But frankly, the guys who think HIV is no big deal at the individual level are not entirely wrong these days, at least in rich, socially tolerant countries with good health systems.

Ken Mayer and his colleagues in Boston end their paper by saying that we need “Innovative programmes that facilitate education and skills building around safer sex when MSM are relatively recently diagnosed”. But frankly, we’re never going to figure out how we should prevent HIV in a post-AIDS world if we can’t make a convincing case to the individuals most at risk that we should prevent HIV.

For more nerdy observations on the Mayer and co. paper, read more.
As far as I can tell, the paper looks at syphilis, gonorrhea and chlamydia in HIV-positive gay men who come to a gay-friendly clinic for health care. It seems that guys are deemed to have had an STI if they test positive in a baseline study screening or if their clinic records say they had any one of those infections in their urethra, rectum or throat in the year before the start of the study or. But the arse or mouth appear to come only from clinic records. If that’s correct, then pharyngeal and rectal infections may be substantially underestimated. Rectal STIs include infectiousness in Bottoms, which is worrisome for negative guys practicing “strategic positioning” — only ever being the Top when they’re barebacking with a guy who is poz and who’s status they don’t know.

The study does have some measures of viral load at baseline; guys with detectable viral load (>75 copies/ml) were 68% more likely than guys with undetectable viral load to have barebacked with someone who might be negative. Since viral load is THE key in transmission, that’s not good. The effect disappears, though, if you sling it in to a model with a lot of things that are related to viral load, including meds and years since diagnosis. But it would be hard to wave too many flags about that anyway — viral load is apparently measured only at baseline, while the unprotected sex we don’t want to see in conjunction with it can be up to six months previously, and STIs (which also lead to spikes in viral load) up to a year previously. As though that’s not enough confusion about time periods, drug use is measured up to three months previously.

Just to be clear, the thing that the Boston gay community (and/or the public health authorities) need to be most worried about in terms of ongoing spread of HIV among guys in care, are, in this order:

• Poz guys with detectable viral load being the top in bareback sex with guys they don’t know are poz

• Poz guys with urethral STIs being the top in bareback sex with guys they don’t know are poz

Drug use: ah yes. I was surprised by two things. The first is that “binge drinking” (defined as five or more drinks in one day any time over the last three months — oh dear, oh dear) is so low — at 19%. The other is that crystal meth use is so high — 23% worship at the shrine of Tina. If there is one mistress that makes you behave worse than most, it is surely Tina.
On the other type of drug, there’s a bit of a surprise too while 66.1% of the nearly 400 guys in this study are on ARVs, only 54.4% had an undetectable viral load. That means that over 11% of these men are on meds and don’t have an undetectable viral load in a single baseline measure. Pause for thought for those who’s prevention strategy relies on thinking “Oh well, if he’s poz he’s probably on meds so he’s not infectious”.

One more deeply curious finding. Men who had any unprotected anal sex with someone that they didn’t know was HIV-infected over the last six months were over four times more likely to have had an STI in the last 12 months (odds ratio 4.42, 95% CI 1.88 – 10.36). But when they looked separately at insertive and receptive anal sex, they found lower chances of infection for both. In receptive anal sex, which you might expect to be associated with greater risk of STIs generally, guys were under four times as likely to have had as STI as those who never took it up the butt without a condom, and despite the fact that it’s by definition a subset of the previous measure, the confidence interval is narrower (OR 3.86, 95% CI 1.78 to 8.28). In insertive anal sex, which you’d expect to be perhaps less likely to associated with STIs, it was lower still: guys were just over twice as likely OR 2.11, 95% CI 1.04 to 4.30). I’m on a flight right now without my trusty stats textbooks, but in my mind, those you can’t have smaller sub-sets of the same measure giving you tighter confidence intervals.

Maybe they are not sub-sets. The “risk behaviour” measure is unprotected anal sex with anyone who is not known to the poz guy to be infected with HIV. In the results section of the paper, the other two measures are described as “unprotected serodiscordant insertive anal sex” and “unprotected serodiscordant receptive anal sex”. In other words, it’s possible that it excludes the “don’t know his status” partners, although the truth of it is that the only status you every really know for any length of time is poz. As the paper says in its introduction, poz guys have higher rates of all these infections, especially syphilis, so if there were a way of restricting the analysis to those who only had truly negative partners, you’d perhaps get those lower rates. But you’d still likely have wider confidence intervals.
In my day job, I teach a course in scientific writing; it includes a fair bit of paper critique work. If I were to add this paper to the course, what else would you expect students to pick out?

Blog comment is divided between “HIV is everyone’s problem”, “HIV is a gay problem” and “HIV is a black problem”. I have the highest respect for DC’s attempts to get to grips with this long-neglected issue, but I was frankly disappointed to hear city HIV director Shannon Hader take the “everyone is at risk” approach in an interview with NPR. So much so, that I went and made a little graph from the data:

I’ve just looked at blacks and whites, I’ve included gay men who inject drugs in the gay men data, and I’ve ignored unknown modes of transmission — the full data are in the table at the end of this post if you’re curious. But essentially, these data show fairly dramatically that in DC HIV affects black people far more than white people, and it affects gay men far more than straight men. A significant number of black women but virutally no white women are infected through drug injection, but more black women are infected because they have sex with black men. That’s the way it is, and saying that there are some infections in every demographic therefor everyone is at risk won’t change that. I should note that these are individual diagnoses, not rates — if we had denominators, the rates among men who have sex with other men and among drug injectors would dwarf heterosexual tranmission.

I was also surprised by NPR’s handwringing over low rates of condom use. Apparently, though the data are not in the report, one in three adults in DC used condoms the last time they had sex. That’s neither low nor high; it’s meaningless. We don’t need people to use condoms every time they have sex. We need people to use condoms with every partner who is remotely likley to be of a different infection status than themselves. As you can see from the graph, lots of potential partners including most white heterosexuals who don’t shoot up drugs and have sex with others like them don’t fall into that category.

One thing that shocked me in the report was the astoundingly high proportion of people who are still testing very, very late (within a year of being diagnosed with clinical AIDS, so when they’ve been walking around with undiagnosed infection for an average of eight or nine years). It’s still an astoundingly high two thirds. It does, however, look like the DC health department’s valiant efforts to increase testing and pick up more cases in settings where they are most likely to be found are bearing fruit in at least one way. The average CD4 count at diagnosis has risen from under 200 in 2004 to 332 last year. The CD4 count is a measure of how damaged your immune system is at the time of diagnosis. If the health system doesn’t catch you and put you on anti-retrovirals until it’s fallen below 200, you’ve got a much lower chance of survival than if treament starts earlier.

So I was especially delighted when Dag Brück sent me this wonderful piece of Word Art. Dag put “The Wisdom of Whores” into Wordle, and here’s the result:

Wisdom World Art: Click for larger version

I love the random “headline”: happy health risk. Better yet, within the giant ARV component, creep the two little words: big caveat. It took me 8,000 sciency words to say this in a paper I sunmitted last week. Art does it in three.

The product in question is Pro2000 gel, and the results of the first large trial on more than 3,000 women were reported yesterday at the Conference on Retroviruses and Opportunistic Infections. CROI is all a scientific conference should be (and all the biannual AIDS Circus is not), and you can see and hear every presentation online. The results of the Pro2000 study show that women using the gel were 30% less likely to become infected with HIV than women using a placebo, and a third less likely than women using nothing at all. The reason the researchers are not screaming about it more joyfully is that the results are “not statistically significant”. Meaning, in this particular case, that we can only be between 90 and 94% sure that the difference in infection rates were really the results of the gel, and not the results of pure chance.

This is just silly. If I told you that there was a 94% chance that the red car was a third more likely to crash than the blue car, which would you drive? Yet we’ve managed to establish a norm in the scientific community that only differences that have a 95% probability of not being due to chance can be trusted. For nerds, that means a “p value” of five percent or less is sacrosanct: (p <0.05) has become a talisman of good science. I’m not the first to remark that things can be significant without being statistically significant — economist Tim Hartford wrote a column on statistical significance and Guinness in the FT only last week.

Someone at the conference remarked that “none of us in this audience worship at the alter of the p value of point oh five” but in fact, many of us do. Another thing that researchers at CROI have been bending over backwards to do is to prove that people on ARVs don’t have more risky sex than people not on ARVs. (Aside: this completely misses the point about “behavioural disinhibition” — jargon for “Oh look! HIV won’t kill me! Let’s party!”. What matters is not so much what infected people do once they are on meds, what matters is what uninfected people do because they no longer see any visible connection between unprotected sex and death. Still, people feel the need to show that ARVs don’t make you screw more.) So when a group working in Uganda showed that people on ARVs were 70% more likely to have an extramarital partner than people not on ARVs, they were happy to worship at the alter of the p value of point oh five. In this case, the p value was 0.09 — in other words there was a greater than 90% chance that the differences were real, but researchers were able to say there were “no differences”. We worship from the underside of the alter, too. A larger study looking at ARVs, risky sex and HIV transmission found that unprotected sex was “significantly lower” in those on ARVs. In fact, 17% of those on ARVs reported unprotected sex compared with 19% of those not on ARVs. The difference may have been statistically significant, yes, but does it meet the most important test of significance, the “So What?” test? Almost certainly not.

Epi-rant over. The microbicide trial (and the fact that there is very low transmission from people on antiretrovirals to their partners in the two ARV studies I’ve just ranted about) wasn’t the only good news at CROI today. Giving monkeys antiretrovirals before exposing them to SHIV rectally worked pretty well, too, which bodes well for PrEP in humans. Disappointingly, though, it worked best when the drugs were given between a week and a day before exposure — ARVs taken just a couple of hours before exposure didn’t have much effect. Bang goes my dream of earning millions with an Ecstasy/ Viagra/ Tenofovir combination pill for big nights out. Maybe I’ll just have to settle down and get a real job.

Thanks to Roger for prodding me to spend my day at a virtual conference…

Interestingly, I can’t find anything anywhere from anyone who has actually nicked the pills off their Mum, crushed them up, rolled them and smoked them. A politician says it feels like taking smack, but his account doesn’t sound very first hand account on either score. Tooli Nhlapo, a documentary maker with SABC, said that after they smoke the meds “The children do not know where they are and they stop making sense”.

How much sense were they making in the first place? Quite a lot, in a teen-eyed view, you might argue. Smack costs money. ARVs are free, to those who need them. It’s just a matter of getting the meds from the hands of patients to those of bored, thrill-seeking teens. More than one in 10 teens is infected with HIV in some parts of the country and bored, thrill-seeking teens are the very ones most likely to be infected. So they could stop swallowing their meds and start smoking them (especially if we press ahead with the WHO’s “potential strategy” of testing everyone annually and putting pills in the hands of every infected person right away). Others teens are apparently buying ARVs off people who would rather have cash to buy booze than take their meds. It slightly begs the question: if teens have cash to spare why don’t they just skip the extra step and get high on booze right away? Are manufacturers of alcopops missing a trick in the South African market?

It’s hard to know how much of this is real and how much is just another silly season media beat-up. I notice that the usually very sensible Treatment Action Campaign doesn’t dignify the reports with any comment. But if the reports are even partly true, it is one more strike at the heart of the prevention approach which relies on young people making sensible decisions about their long-term future in the face of diversions like sex or drugs that will deliver fun right now.

Michael tells the story more patiently than I’d be able to. As several people pointed out, including Dharmawan, Willem and Catherine. Read. Really, please read. What Michael doesn’t say is that Indonesia has pocketed more than US$ 400 million in foreign taxpayers’ money to deal with HIV so far, and has another US$ 130 million from the Global Fund on the table. And it can’t even assure treatment in the big, well-funded, teaching hospitals designated to provide ARVs. No need to point out that treatment is the EASY bit, the big, politically popular success story. People need it because we fail so badly on the cheaper, easier prevention services, like making sure that kids who inject heroin in Indonesia don’t also inject HIV. If it doesn’t make you angry that we’re using your money so badly (if you’re a European or North American tax payer) or serving your needs so poorly (kalau seandenya anda orang Indonesia, apalagi orang Indonesia yang terinfeksi HIV…), this blog’s no place for you.

The paper is based on a mathematical model which assumes that all adults get tested for HIV once a year, and all get put on treatment as soon as they test positive. This “theoretical strategy” could reduce HIV transmission to negligeable levels within 10 years of being implemented, even in an epidemic as widespread as South Africa’s.

This approach, referred to in an earlier galley version of the paper as a “proposed strategy” rather than a “theoretical strategy”, is born of despair. HIV is a (largely) sexually transmitted infection which can be prevented by cutting down on the turnover of sex partners and using condoms. But many people (I’m one of them) would rather risk infection than do what it takes to prevent it. And the reluctance to adopt safer behaviour is particularly pronounced among some of the people who need it most — heterosexuals in Africa and gay men worldwide. The advent of treatment and the disappearance of AIDS appears to be eroding even further the already feeble motivation to cross our legs or use condoms in non-commercial sex. Vaccine research is in a slump, and microbicides continue to disappoint. So what the hell, let’s argue for universal (voluntary — another insistent addition to the paper at the galley stage) testing and treatment.

The problem is that getting all adults to take an annual HIV test and supplying over 30 million people with expensive drugs that have to be taken with daily diligence to reduce the very real threat of resistance is at least as far fetched as persuading people to use condoms. Lesotho’s Know Your Status campaign, which aimed to provide testing and access to treatment for 1.3 million people, racked up just 25,000 tests close to the time it was scheduled to be finished, according to a report from Human Rights Watch. Even in the countries that have very strong health systems where testing is actively promoted among those most at risk we are failing to get people tested and treated. The new data published yesterday by Britain’s Health Protection Agency show that despite a push for more testing, exceptionally high levels of knowledge and awareness and universal access to free treatment, over 3000 gay men were newly diagnosed with HIV in the UK in 2007. One in five weren’t diagnosed until after they effectively had AIDS.

Are current prevention efforts doomed to failure? Perhaps. But that is no reason to replace them with a treatment approach which is just as likely to be doomed to failure. Interestingly two of the paper’s authors also contribute to The Lancet a commentary that questions the feasibility of the “theoretical strategy” that promises such good results. Instead of running mathematical models, perhaps the (rightly) concerned folks at WHO could persuade a single rich, well-governed country with a strong health system and minimal issues of stigma to show their “theoretical strategy” can work in practice. After all, mathematical models show that the theoretical strategies of abstinence, mutual monogamy among the uninfected and universal condom use are 100% effective in wiping out HIV. Though we’ve been pushing those for years, we’ll still have 2.5 million new infections this year.

Now we know. “The £15 pill to protect against HIV” was the headline in Metro, the Daily Mail’s give-away version. More muted than I would expect, but Metro is really the Mail Lite. The article, bylined Jo Steele, does indeed seem to have been stolen wholesale from a much more rigorous piece in the New Scientist, headlined Safer Sex in a Pill.

Missing from the headline is a question mark. While the New Scientist piece does point out that we’re not sure that PrEP works, the Metro piece glosses over the uncertainty, saying (wrongly) that Viread and Truvada “have proven successful in human trials involving 19,000 gay men”. Anthony Fauci is quoted as saying “There’s a lot of buzz about PrEP. There’s some cautious optimism this will work”. I see a lot of optimism, not much of it cautious. My favourite example of the glass being half full comes from Bob Grant, from UCSF. He doesn’t think that people who take a pill so that they can have more unprotected sex will actually have more unprotected sex.

“Bob Grant, who researches HIV prevention at the University of California, San Francisco, and helped run the African trial, speculates that some users may even have less unsafe sex. “PrEP might put people in a different frame of mind,” he says. “When people take a pill a day, that reminds them that they are at risk of catching HIV.”

Hmmm. I took pills every day for about two decades so that I could have unprotected sex without catching pregnancy, and I never once thought: ah yes, I took a pill this morning to help avoid pregnancy, and sex can lead to pregnancy so I’d better not have sex.

I’m all for increasing HIV prevention options, especially ones that have been proven to work (if only in some circumstances, some of the time). But I do think we also have to recognise that what works for an individual may not work so well at a population level — we’ve seen this with the roll-out of ARV among gay men in rich countries. Individually, infectivity goes down, but at the population level, new cases are rising. Look out for a paper and commentary along these lines in The Lancet next week. A number of authors from the WHO will argue that testing all adults annually and putting everyone infected on ARVs would essentially wipe out HIV transmission within 10 years. In a commentary, some other authors from the WHO conclude that it would be a good idea, but might be hard to do in practice.

The truth is that what works in a clinical trial (or even on the much thinner ice of a mathematical model) can not always be made to work in practice. There will be a lot of resistance to PrEP at first, not least from the people who already object to giving out condoms. It’s no bad thing to start thinking now about how to deal with that resistance. (IRMA draws our attention to a Congressional briefing on PrEP on December 4th.) But let’s not get too far ahead of ourselves. There’s no point cashing in political capital in favour of PrEP until we’re sure it works.