A Static Epidemic

New estimates show that public health efforts in the US have had little success controlling the number of new HIV infections over time

By Regina McEnery

Twenty-seven years after the US Centers for Disease Control and Prevention (CDC) published a report about a mysterious cluster of pneumocystis carinii pneumonia cases among five gay men in Los Angeles, the number of people living with HIV/AIDS in the United States has grown to an estimated 1.2 million, according to the most recent figures (see www.cdc.gov). According to the updated prevalence estimate, derived from an improved methodology, of the 1.2 million HIV-infected individuals in the US today, an estimated 34% have AIDS, 42% have not yet progressed to AIDS, and nearly 25% are unaware of their HIV infection (Public Health Rep. 122, 63, 2007).

Morbidity and mortality associated with HIV/AIDS have waned dramatically since the days when AIDS was a virtual death sentence, contributing to the ballooning HIV prevalence. HIV-related deaths in the US have declined significantly since the advent of highly active antiretroviral therapy (HAART) in 1996—plummeting by more than 70% between 1995 and 2004. Once the leading cause of death among Americans ages 24-44 in 1995, HIV is now usually a chronic condition if managed effectively with some combination of the 32 different antiretroviral drugs that act on the virus or its target cells in different ways to limit viral replication (Mon. Vital Stat. Rep. 45, 11, 1997).

But what disconcerts public health researchers is that the latest surveillance data reflect a static epidemic in which the HIV incidence in the US has not changed much since 1994. Despite continued efforts to improve education and promote effective and available interventions like condoms, public health agencies have had little success in controlling the number of new infections.

This worrisome situation will be highlighted in a much-anticipated surveillance report from the CDC that incorporates more comprehensive data from state registries and a more accurate method of identifying recently HIV-infected individuals using the serologic testing algorithm for recent HIV seroconversion (STARHS) technology. The STARHS method employs a combination of assays to draw its conclusion. A standard enzyme immunoassay (EIA), which is used to diagnose HIV infection by the detection of HIV-specific antibodies, is coupled with a less sensitive or “detuned” antibody test called BED HIV-1 Capture EIA. Antibody responses are generally weak soon after an HIV infection occurs, so if HIV-specific antibodies are detectable by the less sensitive assay, researchers conclude that the individual was not recently infected. If antibodies are detectable by the normal EIA assay, but not the less sensitive one, researchers using the STARHS methodology conclude that this individual was recently HIV infected. This model has been increasingly used around the world to estimate incidence in the absence of studies that directly track HIV incidence by following cohorts of uninfected individuals over time.

The new HIV incidence figures calculated using this method were submitted for publication in an academic journal by the CDC last year to make sure the methodology, emerging data, and conclusions were scientifically rigorous. The agency says the data is still undergoing review. The new incidence estimates are widely expected to be announced sometime this year and are expected to show that the number of new HIV infections for 2006 are significantly higher—perhaps by as much as 20,000 infections—than the annual estimate of 40,000 HIV infections per year often cited by public health departments to describe the steady pace of the epidemic in the US since 1994. While there is already plenty of speculation surrounding what the revised HIV estimates actually mean—discussions fueled mainly by the agency’s slow pace in releasing the data—those familiar with the new methodology say the more accurate and timely epidemiological data probably won’t be portrayed by the CDC as a major resurgence in overall incidence.

“Most likely it is just an upward adjustment and a more accurate estimate of what has been occurring in the last decade,” says Walt Senterfitt, a California epidemiologist involved with Community HIV/AIDS Mobilization Project (CHAMP), a national alliance of prevention activists.

What the new incidence estimates will dramatize is how little progress the US has made in preventing the spread of new HIV infections among adults, particularly within high-risk populations. The updated incidence data should provide a much clearer picture of where the epidemic is heading in the US, helping to identify populations and individuals at highest risk of HIV infection. This could eventually offer researchers conducting clinical trials for vaccines, microbicides, and other biomedical interventions more reliable incidence estimates to use when designing efficacy trials.

Vaccinations in the Phase IIb STEP trial, which had enrolled 3,000 men and women in North and South America, the Caribbean, and Australia, were stopped in September after Merck’s adenovirus serotype 5 (Ad5)-based candidate, MRKAd5, showed no protection against infection (see A STEP back?, IAVI Report, Sept.-Dec., 2007). The majority of volunteers in this trial were men who have sex with men (MSM), but 1,100 were also women at high risk of HIV infection, 473 of them at US sites. During the trial only one female volunteer in either the vaccine or placebo group became HIV infected. Investigators involved in the STEP trial initially speculated that perhaps the women enrolled were not at particularly high risk of HIV infection, but Susan Buchbinder, a principal investigator for the STEP trial, said recently at the Keystone HIV Vaccines Symposium that there were high pregnancy rates among the female volunteers, corroborating the fact that they were having unprotected sex. She said that the low HIV incidence in women during the trial was more likely due to the lower HIV prevalence among heterosexual men in the US as compared to MSM.

More accurate incidence data will also help eliminate other problems with clinical trial design. After analyzing several non-vaccine HIV prevention trials in preparation for an Institute of Medicine (IOM) Report on the methodological challenges of conducting such trials, Harvard University biostatistics professor Stephen Lagakos and IOM Senior Program Officer Alicia Gable concluded that clearing up design deficiencies, including basing trials on unreliable incidence data, is key to overcoming the kinds of problems that led to premature termination of some recent late-stage non-vaccine HIV prevention studies (N. Engl. J. Med. 358, 1543, 2008).

Readjusted incidence numbers

Harold Jaffe, who headed up the CDC’s National Center for HIV, Sexually Transmitted Disease (STD), and Tuberculosis Prevention when the Atlanta agency began developing its new HIV monitoring system, considers the system a more accurate way of calculating HIV incidence than the previous methods of back-calculation and data synthesis. Back-calculation looked at the number of individuals currently diagnosed with AIDS, then moved backward in time to estimate when and how many HIV infections would have had to occur per year to add up to current AIDS caseloads.

“Until HAART became available in the late 1990s, we believed it was possible to make reasonably accurate incidence estimates using back-calculation,” says Jaffe. “Once treatment became available widely, that method became unreliable. We decided that we needed a new system,” adds Jaffe. Once effective therapies altered the incubation of HIV, the CDC moved on to data synthesis, which generates HIV incidence estimates using the size of the populations at risk and the relationship between HIV prevalence and incidence. Data synthesis was used by the CDC in 1994 to determine that there were about 40,000 new HIV infections occurring in the US each year. This incidence estimate has essentially been used ever since.

The CDC expanded its existing case surveillance system several years ago to include STARHS to try and tease out recent HIV infections from longstanding ones in the population. “Fundamentally, it’s a much better approach than what we had before and it’s an advance in estimating incidence,” says Jaffe. But the STARHS methodology also has its critics. Three years ago, the Joint United Nations Programme on HIV/AIDS (UNAIDS) recommended that countries not use the BED or detuned assay because it appeared to have overestimated HIV incidence in some African countries and Thailand. But in early studies validating its use on subtype B infections from US cohorts with known rates of seroconversion, the BED assay performed well, says Jaffe.

In either case, Jaffe predicts the new incidence estimates will stir controversy among those who feel AIDS prevention dollars are being squandered, as well as those who believe efforts are underfunded. About 4% of the US$23.3 billion allocated by the government in fiscal year 2008 to fight HIV/AIDS globally was spent on domestic prevention efforts, according to a Kaiser Family Foundation analysis.

Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, says there is no single reason for why the US has hit a “brick wall” in reducing the number of new HIV infections. He says AIDS has lost the terrifying persona that once served as a powerful incentive for careful behavior, while poverty, substance abuse, homophobia, and poor healthcare continue to put a disproportionate percentage of African Americans at risk for HIV.

“This makes it even more compelling for us to find a vaccine for HIV,” says Fauci. “It is needed universally and we have reached a point, particularly in the US, where we can’t get beyond the 40,000 new infections a year.”

Because most HIV infections in the US today are transmitted sexually or by the sharing of used needles among injection-drug users, prevention efforts in the US have tended to center around condom promotion and distribution, needle exchange, HIV counseling and testing within high-risk communities, and sex education, including abstinence-only campaigns. Some of these interventions, notably syringe exchange, appear to have helped to reduce transmission of HIV, the latest data shows. The share of new AIDS diagnoses attributable to injection drug use dropped from a high of 31% in 1993 to 18% in 2006, and although the methodologies used in evaluating syringe-exchange programs are subject to a number of limitations, these studies generally have found that needle exchange has had a major hand in reducing HIV transmission within this high-risk group, CDC surveillance data from 2006 shows. There are currently 185 known syringe-exchange programs operating in 36 states, according to the CDC.

But many of the behavioral interventions launched by state and local health agencies, grass-roots organizations, and faith-based groups over the years have not been well-studied, and epidemiologists and social scientists tracking the epidemic tend to think the approaches have had minimal, if any, effect in reducing infections within communities shouldering the biggest burden of HIV/AIDS—men who have sex with men (MSM) and African Americans.

“I think the reality is that HIV prevention through behavioral change, which is what we have available for adults, isn’t that effective,” says Jaffe, now at Oxford University. “Fundamentally, it’s hard to change human behavior. When you look at published studies on prevention techniques, they have been done on a small scale. It is difficult to say how well they would work in the general population. I think we need to be asking harder questions.”

From 2003-2006, the most recent period for which data is available, the estimated number of HIV/AIDS cases in the US increased among MSM but remained stable among adults and adolescents who contracted HIV through high-risk heterosexual contact, according to the 2006 HIV/AIDS Surveillance Report. MSM and persons exposed through high-risk heterosexual contact accounted for 82% of all HIV/AIDS cases diagnosed in 2006, the CDC report says, basing its estimates on data collected from 33 states and five US-dependent areas that have had confidential name-based HIV reporting since at least 2003 (see Figure 2). Confidential name-based reports include data on patient demographics, HIV risk behaviors, laboratory and clinical events, and virologic and immunologic status. State and local health departments collect the data and forward it to the CDC, minus the patient’s name and other personal identifiers.

Figure 2. Shift in HIV transmission patterns in the US based on estimates from the Centers for Disease Control and Prevention

View Larger Image

Although the estimated number of newly-diagnosed HIV/AIDS cases remained stable among blacks and Latinos while increasing among whites, blacks still accounted for 49% of all HIV/AIDS cases in 2006, the CDC surveillance report found.

Women represented 26% of HIV/AIDS cases diagnosed in 2006, compared to just 8% in 1985, and the CDC estimates there are about 300,000 women living with HIV/AIDS in the US today. Black women accounted for two thirds of new AIDS cases among all women in 2006, according to the CDC surveillance report.

Though perinatal transmission has declined dramatically in the US since the start of the epidemic, mostly because of the delivery of prompt antiretroviral therapy to pregnant women and their babies, there were still 609 infants who contracted HIV before or during birth from their mothers during the years 2002-2006, the CDC reports. The CDC recommends HIV screening during prenatal visits and five states even mandate it, but hundreds of infants still slip through the cracks because so many women are becoming newly infected with HIV every year.

“Mother-to-child transmission remains a continued problem because of the several hundred thousand women infected with HIV,” says Dr. James Curran, of Emory University, an epidemiologist who investigated the very first cluster of AIDS cases in 1981 for the CDC. “We have testing during pregnancy and rapid use of antiretrovirals and other mechanisms. And it still is not eliminated.”

Infected and undetected

The difficulties in reducing HIV in the US have been underscored by another statistic that public health agencies believe is partly to blame for the static rates of transmission. Using back-calculation methods, the CDC estimates about 25% of the 1.2 million individuals living with HIV/AIDS are unaware that they are infected. Because some of the people with unrecognized HIV infection may transmit the virus unknowingly, perhaps for years because of the virus’ long latency period, the CDC expanded its routine testing recommendations two years ago to include all adolescents and adults ages 13-64, rather than just those in high-risk groups.

It is unclear as of yet whether this recommendation will help identify infected individuals, providing them with earlier access to treatment and care services and possibly lowering the chances that they will transmit the virus to others. State and local health departments support the new testing guidelines for physicians, clinics, and hospitals even as they struggle for ways to implement the new guidelines and the burden of caring for the additional cases that will surely surface.

“It’s just like blood pressure or diabetes, if you don’t know you have a problem you can’t get help,” says Tom Liberti of the Florida Department of Public Health, which tested 300,000 individuals last year, more than any other state.

Whatever long-term impact the testing guidelines will have in altering the status of the US epidemic, tracking it continues to be a complex epidemiological exercise that, paradoxically, seems to grow more difficult as public health agencies become more skilled at collecting and analyzing data. Compounding the confusion has been the mosaic of surveillance systems adopted by states since the start of the epidemic. Some have used confidential name-based systems, others have favored code-based reporting that replace the patient’s name with their initials, date of birth, or other secret identifier and still others have used a hybrid system.

It took 21 years, for instance, for all states and dependent areas to implement HIV case reporting—the last ones just started in 2004. And it wasn’t until 2005 that the CDC recommended all states and dependent areas adopt confidential name-based HIV infection reporting to “better monitor the scope of the epidemic.” States are finally on board, but it will be at least three years before the CDC is able to establish trends, particularly at the state level.

Curran says HIV incidence is also hard to determine in the US because compared to AIDS-ravaged areas like sub-Saharan Africa, the incidence in the US is fairly low and the epidemic is not equally distributed across the geographic population. “When you have unequal distribution, it becomes even more difficult to find out what the incidence is,” says Curran. “For example, a household survey would under-represent many of the people at high risk for HIV.”

While measuring HIV incidence is very hard, getting it lower remains the goal, Curran says. “Each year, new people become at risk and the availability of good therapies has shifted attention away from the horrible death rates,” says Curran. “So we now have a whole generation of young people who have grown up without having AIDS be so visible. Those people have a different feeling of risk.”

Social scientists and epidemiologists who have studied the epidemic over the years aren’t entirely sure what’s driving the racial disparities in HIV infection among Americans, particularly black and white MSM. A meta-analysis published last June of 53 different studies found behavioral risk factors did not appear to solely explain the elevated HIV rates among black MSMs (AIDS 21, 2083, 2007).

Greg Millett, the CDC behavioral scientist who led the study, says lower rates of antiretroviral use and higher rates of undiagnosed HIV infection among black men might explain the disparity, or the fact that the occurrence of STDs was higher among black MSM. “But we just don’t know,” says Millett. “The fact that the epidemics might be different in sub-populations of MSM has not really been looked at.”

Ron Stall, a behavioral scientist at the University of Pittsburgh, evaluated US incidence data from 1995 to 2005 and estimated that slightly more than 2% of MSM in the United States were becoming infected every year. Stall, who presented his results in February at the 15th Conference on Retroviruses and Opportunistic Infections, said HIV prevention needs to incorporate multiple mechanisms—from better access to healthcare and treatment for co-morbid conditions and policies that promote gay men’s health, to individual interventions such as condom distribution—to change the course of the epidemic in the MSM community.

“The risk levels among gay men are driven partly by the behavioral risks,” says Stall. But he also notes that high incidence within some communities is also driven just by the sheer fact that HIV prevalence is so high. “Just by being a human being in these [areas with] very high prevalence rates, over the long haul the chances of being infected are pretty good.”

Adaora Adimora, a professor of medicine at the University of North Carolina who has studied HIV rates in the African American community, particularly African American women, says there are many issues that need to be addressed for HIV prevention to work. These include the high incarceration rates among black men and the lack of services offered to them upon release from prison, high rates of STDs among African Americans, inadequate healthcare, and other factors that have put an extraordinarily high number of black women at risk for HIV. Studies have also found that fewer HIV-infected black MSM were likely to be on HAART, despite markedly higher HIV prevalence (AIDS 21, 2083, 2007).

“You really have to regard this as the truly urgent situation that it is,” says Adimora. “HIV prevention is going to require a fresh look.” She says behavioral interventions are a good idea but fall short. “To the extent that factors outside the individual are playing a significant role in driving HIV incidence, condom use alone won’t be enough,” Adimora says.

AIDS advocates frustrated by the failures in controlling the US epidemic want a national AIDS strategy that incorporates more money for prevention, more rigorous studies of existing prevention methods, and better access to healthcare. “There is a mindset out there now that we will never have behavioral or social strategies that work,” says Julie Davids, executive director of CHAMP. “No, we are not going to have a magic bullet, but we need to have a combination of approaches that could be rooted in a biomedical intervention.” Biomedical interventions could include a preventive vaccine, microbicide, or pre-exposure prophylaxis for HIV.

With the cost of treating AIDS growing yearly in the US, advocates are also increasingly worried about how state and local governments, which shoulder most of the cost, will be able to afford programs over the long haul. In the US, HIV/AIDS spending through Medicaid—a government health insurance program that is one of the primary funders of treatment for HIV-infected individuals—increased from $1.3 billion in 1994 to $6.3 billion in 2006, according to a report by the Kaiser Family Foundation. This has reinforced the need for effective HIV prevention strategies.

But those on the frontlines of the US epidemic say a big challenge is keeping AIDS awareness and prevention on the radar screen. Advocacy for HIV prevention has never reached the same level as treatment advocacy in the US. Scores of activists terrified by the specter of AIDS in the 1980s and 1990s fought to save the sick and dying, and had little time to take care of the healthy, says long-time AIDS activist and playwright Larry Kramer.

Kramer, who helped establish the landmark HIV advocacy and activist groups Gay Men’s Health Crisis in 1982 and AIDS Coalition to Unleash Power (ACT UP) in 1987, says many activists also found AIDS vaccine science confusing and incomprehensible. He says the government has done a poor job communicating the scope of the research, while the failure of some vaccine candidates hasn’t helped. “It never caught on, even until this day,” says Kramer. “It has always been a mysterious bunch of hocus-pocus.”

Fauci, whose long career in AIDS research and public service put him in the center of the discovery of HIV/AIDS, the fight for treatment and its profound impact on millions of lives, and now the continuing quest to develop an AIDS vaccine, says in some ways the US has become a victim of its success. “Since we have good therapies and people are living normal lives, the perception of what a risk behavior might ultimately mean is different than what it was,” says Fauci. But in an editorial he authored recently in Science magazine, marking the discovery of HIV as the causative agent of AIDS 25 years ago, he said, “New infections far outstrip our ability to treat everyone infected with the virus: around three people are newly infected for every person put on therapy—and current HIV therapy is a life-long commitment.” He called the discovery of a safe and effective HIV vaccine “our best hope for ultimately ending the pandemic.”