The HIV prevention story, however, hasn't been as happy. After a sharp drop in the late 1980s, the number of new infections in the U.S. reached a plateau, and holds steady at around 56,000 per year. Globally, there were an estimated 2.7 million new infections in 2008, and for every two people who start treatment, three more become HIV positive. Once-promising prevention tools such as vaccines and microbicides have seen very slow progress.
But what if HIV treatment itself is the best weapon against new infections? HIV medications dramatically lower the risk of transmitting the virus, and some experts think that if every person with HIV started treatment, transmission of the virus could finally be halted.
It is well known that lowering HIV viral load decreases the likelihood of transmission. This is already the standard approach for preventing mother-to-child transmission. Giving drugs like AZT or Viramune -- or even better, a combination regimen -- to an HIV-positive woman during pregnancy and delivery, and to the newborn immediately after birth, reduces the risk of the baby becoming infected to less than 2%. There's good reason to think a similar strategy would work for HIV transmission via sex or sharing drug injection equipment.
Early in 2008, the Swiss Federal Commission for HIV/AIDS set off a vigorous debate when it issued a report stating that a person taking HIV treatment with an undetectable viral load for at least six months and no other sexually transmitted diseases essentially cannot transmit HIV through heterosexual contact. The statement did not apply to gay men, who have not been studied as thoroughly as heterosexual couples.
At the 2008 International AIDS Conference in Mexico City, Commission President Pietro Vernazza explained that the statement was intended to aid Swiss physicians in discussing sexual risk with their patients, not as a universal recommendation. But he reaffirmed that unprotected sex with an HIV-positive person with undetectable viral load is as safe as using a condom -- not 100%, but within a "comfortable range" most people can live with.
This conclusion was based on several studies showing a very low likelihood of transmission between steady heterosexual partners. A study of nearly 400 mixed-status couples in Spain, for example, saw no new infections due to unprotected sex if the HIV-positive partner was taking treatment. A smaller Brazilian study found that transmission only occurred when the positive partner was not adherent to HIV therapy.
In June, the Partners in Prevention team reported in The Lancet that treatment reduced the risk of transmission between heterosexual partners by 92%. This study of more than 3,000 couples in Africa found that the transmission rate was very low (0.37 per 100 person-years). In contrast, 70% of new infections happened when the positive partner was not taking HIV meds and had a viral load over 50,000.
More cautious experts warn that, while HIV transmission may be rare when the positive partner is taking HIV drugs, it is not impossible. In the African study, there was one transmission from a positive partner taking treatment, and some other large surveys have also seen a very small number of similar cases. This may happen because even individuals on effective therapy can have viral load "blips," or brief increases. And a small proportion of people with an undetectable viral load in their blood can have persistent virus in their semen or vaginal secretions.
A concern is that people who think they cannot transmit HIV because their viral load is undetectable may dispense with safer sex or safer needle-use practices. If enough people do so, new infections could theoretically increase, even with a small risk of transmission.
Beyond prevention of individual infections, public health is a key issue underlying a more controversial approach known as "test-and-treat" or testing and linkage to care (TLC). The idea behind this strategy is that increasing the number of people who are taking HIV treatment lowers the "community viral load."
Viral load typically rises to a high level soon after infection and for the next few months. One study in Quebec found that as many as half of all HIV transmissions may happen during this time. Initial HIV infection often causes no symptoms -- or symptoms easily mistaken for the flu -- so most people do not know they are infected until they take an HIV test some time later.
Test-and-treat relies on regular HIV testing to catch each infection as soon as possible, followed by immediate treatment to minimize the time viral load is high enough to increase the risk of transmission. But everyone may not need treatment this early. Traditionally, explains Wafaa El-Sadr from Columbia University, it was assumed that after people acquire HIV they have "many, many years when they're infected but they're actually doing well, or they appear to be doing well."
But test-and-treat says don't wait to start treatment, and not just for public health reasons. While the benefits of early treatment for preventing HIV transmission are obvious, proponents argue that it has other advantages as well. A growing body of evidence shows that HIV causes persistent immune activation and inflammation that can contribute to problems throughout the body -- such as cardiovascular disease and brain impairment -- long before CD4 cells fall into the danger zone. What's more, today's meds are more convenient and more tolerable than earlier drugs and they can keep viral load low enough to stave off resistance.
In late 2009, a panel of HIV experts raised the U.S. guidelines for starting treatment from 350 to 500 CD4 cells; even above this level, half the panel favored treatment and half considered it "optional." Using a test-and-treat strategy, there would be no upper limit.
On the other side, people who oppose test-and-treat -- or who think it's premature -- say we really don't know how well the drugs will work and what kind of toxicities they might cause over the long term.
This poses a potential trade-off between individual and collective benefit. Why should people who do not yet need treatment for their own health take the risk of starting immediately to lower the odds of others becoming infected? HIV therapy is a lifelong commitment, and it may be hard for people to stick to it if they are basically healthy -- and less-than-perfect adherence can lead to drug resistance.
Some HIV advocates have additional reservations. Testing HIV positive and taking HIV drugs can lead to stigma, unwanted disclosure, problems with health insurance, legal consequences (especially around criminalization of HIV transmission), and even domestic violence. Further, there is concern about the possibility of mandatory testing or coercive treatment for the sake of public health goals.
The first support for test-and-treat as a public health strategy came from mathematical models, which plug different variables into equations to crank out predictions. At the 2006 International AIDS Conference, Julio Montaner and colleagues from the British Columbia Centre for Excellence in HIV/AIDS reported results from a "prevention-centered" model that looked at what would happen if all people with HIV worldwide started treatment, leading to no further infections thereafter.
The model showed that within 45 years, widespread treatment could reduce HIV transmission by 70-fold -- from more than 7 cases per 1,000 people to less than 0.1 cases -- and reduce the total number of people living with HIV from 38 million to less than 1 million. Assuming a generic regimen costing $1 per day, the total price tag would be $338 billion. That's a bargain compared to the $650 billion cost of the "current uptake" model, which assumed treatment coverage in low- and middle-income countries would reach the current level in wealthy countries.
"Although treating 100% of HIV-infected individuals worldwide might not be feasible or even ethically acceptable at this time, given the state of the pandemic, consideration of this possibility is worthwhile," the researchers concluded.
Taking financial and ethical barriers into account, Montaner's team devised a new mathematical model focusing on HIV-positive people in British Columbia who need treatment for their own health. If everyone started therapy when their CD4 count reached 350, they calculated that about 2/3 of new infections might be prevented by 2030. If just 75% started therapy, new infections would still decrease by 40%.
Reuben Granich and colleagues from the World Health Organization (WHO) presented another mathematical model in the January 3, 2009 issue of The Lancet. The researchers asked what would happen if all people age 15 and older in a high-prevalence area were voluntarily tested for HIV each year and if all those found to be positive started HIV treatment regardless of their CD4 count.
Using South Africa as a test case, and assuming that all transmission was via heterosexual sex, they estimated that a universal test-and-treat strategy could reduce new infections by about 95%, from 15-20 cases per 1,000 persons per year, to just one case, within ten years. Within 50 years, total HIV prevalence would fall to less than 1% -- essentially halting the epidemic. While the test-and-treat strategy would cost more up front, by 2032 its price tag would equal that of the current approach of limited testing and medically indicated treatment.
"Although other prevention strategies, alone or in combination, could substantially reduce HIV incidence," the WHO team concluded, "our model suggests that only universal voluntary testing and immediate initiation of antiretroviral drugs could reduce transmission to the point at which elimination might be feasible by 2020 for a generalized epidemic, such as that in South Africa."
At the 2010 Retrovirus conference, Moupali Das-Douglas from the San Francisco Department of Public Health reported that a 40% decrease in average community viral load between 2004 and 2008 -- from about 24,000 to 15,000 copies -- was associated with a 50% reduction in new HIV diagnoses (both recent infections and newly diagnosed existing infections).
By 2008, about 80% of newly diagnosed individuals were linked to care, about 90% of them started HIV treatment, and 75% achieved an undetectable viral load. The researchers concluded that these findings "support the hypothesis that wide-scale early [HIV therapy] can have a preventive effect at a population level." Community viral load can serve as a "virometer," or a way to "take the temperature" of a community, Das-Douglas explained. "This helps us see how well treatment is working, but also how well prevention is working, so we can target interventions to those at highest risk."
Julio Montaner's team has also seen evidence that their mathematical models reflect real-life outcomes in British Columbia, where the epidemic is largely driven by injection drug users in Vancouver's Downtown East-side district. In 2009 they reported changes in community viral load and new HIV infections in a group of 2,000 injection drug users. The median community viral load declined steadily, from about 55,000 copies in 1996 (when less than 10% were taking treatment) to about 8,000 copies in 1998, to below 1,000 copies in 2000. And community viral load was a stronger predictor of new HIV infections than homelessness, syringe sharing, or unprotected sex.
Montaner's team later found that between 2004 and 2009 -- a period that saw stepped-up efforts to expand HIV treatment for drug users -- community viral load decreased and the proportion of HIV-positive people with undetectable virus rose from about 40% to about 75%. During the same period, new HIV diagnoses among injection drug users declined steeply, from 150 in 2004 to 80 in 2009. They reported that between 1996 and 2009, the number of people taking HIV drugs in British Columbia increased by 547%, while new HIV diagnoses fell by 52%. For every 100 additional people who started treatment, they calculated that new diagnoses fell by 3%.
In the wake of these findings, British Columbia has invested nearly $50 million in a four-year pilot project called "Seek and Treat" that will try to expand access to HIV treatment for "hard-to-reach" populations including injection drug users, sex workers, and men who have sex with men.
"The more people you put on therapy, the less transmission there is," said Anthony Fauci, Director of the U.S. National Institute of Allergy and Infectious Diseases (NIAID), which helped fund the study. The decrease in new cases "likely could not be explained by anything else." He named test-and-treat as one component of a three-part strategy for controlling the HIV epidemic, along with pre-exposure prevention and finding a functional cure. The National HIV/AIDS Strategy, released this past July, prominently features a TLC approach.
NIAID has launched a pilot study of TLC in Washington, D.C., and the Bronx, New York -- among the highest HIV prevalence cities in the country. Rachel Walensky and the Cost-Effectiveness of Preventing AIDS Complications (CEPAC) team recently reported findings from a mathematical model that estimated that universal testing in Washington, D.C., followed by immediate HIV treatment could increase life expectancy by one to two years and reduce transmission by 15% to 27%. "Test-and-treat will save lives, but it won't stop the HIV epidemic in its tracks all by itself," Walensky said. "It is only a single new and important page in the HIV-prevention playbook."
San Francisco has taken a different route with a new city-wide policy, announced this past spring, to offer treatment to everyone with HIV, regardless of CD4 cell count. "Based on accumulated data we believe all HIV-infected people should be treated with antiretroviral therapy unless there is a strong reason not to," said Department of Public Health Director Mitch Katz. But, he emphasized, "we don't dictate medical practice by policy," and the ultimate decision will remain with patients and their providers.
Public health officials acknowledge that widespread early treatment will likely drive down community viral load and prevent new infections, but the Positive Health Program doctors at San Francisco General Hospital who spearheaded the change said they were primarily motivated by benefits for individuals.
At a packed forum in April, Steven Deeks, whose team at the University of California San Francisco has pioneered research on HIV and inflammation, explained the rationale behind the policy shift. "We should perhaps think of AIDS as acquired inflammatory disease syndrome," he suggested. "Five years ago we said drugs are no fun and we should wait, but the consequence of waiting is that people develop irreversible harm to the immune system. The new paradigm is that while today's drugs are not totally benign, they are less toxic than the virus," he continued. "If I'm wrong, we'll start people [using treatment] a couple years earlier than we otherwise would. But if I'm right and we don't start early, there's no going back."
The policy has generated considerable controversy. San Francisco-based HIV advocacy group Project Inform issued a statement supporting the city's new policy. Other advocacy groups, including New York's Treatment Action Group, decided not to take a position on test-and-treat and earlier therapy until more information is available about long-term risks and benefits.
A recently opened international trial called START -- Strategic Timing of Antiretroviral Treatment -- is designed to provide such data. HIV-positive people with a CD4 count above 500 will be randomly assigned to either start combination therapy right away or delay treatment until their CD4 count falls to 350. The study will look at outcomes like death, progression to AIDS, and serious non-AIDS conditions. But START is not expected to produce results until around 2015, and some public health officials don't want to wait.
On the other side, test-and-treat skeptics express concerns about unknown long-term drug toxicities, drug resistance, and of course the added cost of providing treatment for people whose immune systems are still working well. Some have even implied that drug company profits were motivating the push for expanded testing and early treatment.
"The seeming simplicity of preventing new infections through the [test-and-treat] approach is appealing. But that simplicity hides deeply disturbing truths, including that many people coerced into unnecessary treatment will suffer side effects and treatment-induced diseases," POZ magazine founder Sean Strub wrote on his blog. "It is unethical and irresponsible to coerce or encourage people who are not recommended for treatment under the guidelines to start therapy without fully informing them of the risks."
"It is an experiment based on a mathematical model," he continued. "It is as though a decision has been made to redirect the country's public health response [to] AIDS from proven behavioral interventions, like condoms and prevention education, to the use of antiretroviral medications. Public health officials' focus on treatment as prevention and the pharmaceutical industry's incentive to expand markets are now in dangerously perfect alignment."
To address such concerns, Project Inform clarified its statement, emphasizing support for the more comprehensive approach dubbed Testing and Linkage to Care Plus (TLC+). The revised policy urges people to find out if they are HIV positive and if so, to get into care to address all the medical, psychological, and social issues they may face. Only then would they make a decision with their doctors about whether they are ready for and could benefit from HIV treatment.
"TLC+ is firmly rooted in principles of informed choice by HIV-positive people regarding all aspects of their care, particularly decisions about whether to be tested and when to start HIV treatment," the statement said. "Project Inform views it as ethical and empowering of people with HIV to describe treatment as a possible support for prevention. We believe that most HIV-positive people are altruistic and willing to factor the possible benefits for prevention into their treatment decisions."
While the debate continues, the idea of treatment as prevention has already been widely embraced by medical experts and public health officials. On a global level, WHO announced last year that scaling up access to HIV therapy is important for both individual clinical benefit and for prevention of new infections.
UNAIDS included treatment as prevention as part of its five-pronged "Treatment 2.0" strategy announced this summer in Vienna. The new approach could prevent up to a million new HIV infections per year if therapy were provided for all people who need it under the latest WHO guidelines, tripling coverage from the current 5 million people to 15 million.
Most experts are hesitant to predict that universal testing and treatment could completely halt the epidemic, but many think it would help bring it under control.
"Both syphilis and tuberculosis were pandemic at the end of the nineteenth century, and both epidemics were controlled by effective diagnosis and treatment," pioneer AIDS researcher William Haseltine wrote in a 2009 essay in The Atlantic. "So, too, might the current HIV/AIDS pandemic be slowed until vaccines are someday available."
Over the past 20 years HIV treatment has seen great advances, but much remains to be done, said Montaner. "I hope Vienna will go down in the history of the epidemic as where we concur that treatment is not just good for patients but also for public health, because treatment is prevention."
Liz Highleyman is a freelance medical writer and journalist based in San Francisco.