CDC Explains and Defends Molecular Surveillance System

Alexandra Oster, M.D., discusses the Centers for Disease Control and Prevention's molecular surveillance program at CROI 2019.
Kenyon Farrow

The Centers for Disease Control and Prevention (CDC) used part of a plenary session at the March 4-7 Conference on Retroviruses and Opportunistic Infections (CROI) in Seattle to explain and defend the agency's molecular surveillance program. The presentation was given by Alexandra Oster, M.D., a medical epidemiologist who has worked on this research for the CDC for over ten years.

Oster began her presentation by arguing that cluster detection and response can help bring the nation closer to ending the HIV epidemic by allowing us to identify when HIV is being transmitted rapidly, and when public health can step in to stop transmission.

Collecting Data and Identifying Clusters

The data come from drug resistance testing, which is used when a person enters care or changes their treatment regimen. Health care providers and laboratories collect genetic sequence data and report it to state and local health departments, which then report de-identified data to the CDC. Genetic HIV data have been collected by some health departments across the country as early as 1997, originally as a way to track drug resistance. The program expanded to 27 health departments in 2013, with a broader focus on tracking transmission patterns. It was expanded again last year to all health departments, and the focus is now on cluster identification.

These data have immediate use to patients, because they can help providers determine what antiretroviral drug regimen will work best. When analyzed as a whole, these data can also help public health officials determine whether drug resistance, including resistance to pre-exposure prophylaxis (PrEP), is spreading. But these sequence data can also be used to compare various strains of the virus to each other and identify groups, or clusters, of people with HIV who have similar strains of the virus.

HIV is constantly evolving in a person's body. Right after transmission, the genetic sequence of the HIV strain in the recipient will be nearly identical to those strains found in the person transmitting the virus. Therefore, people whose viral strains are genetically similar may be closely linked by direct or indirect transmission. Groups of closely linked people are referred to as molecular clusters. Identifying a cluster often signifies ongoing rapid transmission of HIV.

In 2016, the CDC established a routine approach that focuses on recent and rapid transmissions and uses a tight genetic distance threshold (meaning the genetic sequences of the virus are very similar). Oster explained that on average in the United States, there are four new HIV transmissions per 100 people living with HIV, or four transmissions per 100 person-years. Clusters that are identified using this method have a much higher rate of transmission. The first 13 clusters identified by the CDC had a rate of 33 transmissions per 100 person-years, and the first 60 had a rate of 44 per 100 person-years.

None of these clusters were identified by traditional epidemiological means alone, which demonstrates that it can be hard to find areas of rapid transmission without molecular data. Traditional epidemiological approaches rely on observing an increase in diagnoses, talking to patients, and locating partners. This can be especially challenging with HIV, because of the time between contracting the virus and diagnosis (the median gap is three years).

Oster pointed out that molecular outbreak detection is not new, and it's not exclusive to HIV. The CDC uses a system called PulseNet that analyzes food-borne illness using molecular data. It is estimated that this system averts 250,000 cases of salmonella each year. And more than 75% of tuberculosis outbreaks are detected using molecular data.

Public Health Response to Molecular Clusters

The ultimate goal is not just to identify clusters but to respond to them. Oster explained that because HIV is transmitted through networks of people, a cluster essentially indicates gaps in prevention services that need be addressed. Any given network is likely to include people who have been identified by health departments because they've been diagnosed, but also people who have yet to be reached by prevention or treatment efforts, because they are undiagnosed or HIV negative but at high risk. Identifying molecular clusters can help public-health officials find these at-risk people.

The CDC, Oster explained, assists health departments across the country in identifying clusters, but it is the departments that have the authority and the responsibility to respond with prevention interventions. She shared three stories in which molecular cluster identification helped health departments respond appropriately.

The first example was a molecular cluster identified in San Antonio, Texas that included 24 Latino men. A review of partner services data found an additional 87 people who were in the network as sexual or needle-sharing partners. This helped the department recognize missed opportunities in which these individuals could have been diagnosed. The investigation also revealed issues with access to PrEP. The city sent out provider alerts in an attempt to address the missed opportunities and reallocated health department funds to scale up PrEP.

In another example, the CDC identified a molecular cluster in Michigan. Everyone in the cluster was virally suppressed, but the cluster was continuing to grow, which suggested that there were undiagnosed individuals in the network. In response, HIV providers worked with patients to identify others in their social network who might be at risk and encouraged increased access to testing, treatment, and PrEP.

A small town in Maryland also took advantage of molecular data when the health department reported six diagnoses in a short period of time. The town normally had about one person diagnosed with HIV per year. The genetic data, however, showed that the viruses were unrelated, other than two individuals who had named each other as partners. Understanding that this was not an outbreak helped the health department avoid overreacting.

Addressing Community Concerns

Toward the end of the presentation, Oster acknowledged the concerns that the HIV community has about the CDC's molecular surveillance system, beginning with the concern over existing criminalization laws. She argued that these laws often criminalize lack of disclosure and potential exposure without diagnosis and regardless of viral load. Oster also noted that as of now, the data cannot assess directionality of transmission and emphasized that the CDC does not care about that.

"Regardless, I want to state that for public health purposes, the goal is not to identify direction of transmission, but rather to detect networks in need of prevention and care services," she said.

Given the concerns, however, CDC is re-emphasizing efforts to protect public health data from non–public-health use. Though Oster did not go into detail on any of these plans, she did say the CDC is requiring all health departments to further strengthen data protection to ensure that the collected data is only used for public-health purposes.

The presentation also noted that communities are concerned about increasing the stigma around HIV transmission. Oster said that the CDC is working to avoid further marginalizing any population. The goal is to act on cluster data in a way that encourages people to seek lifesaving services and to make sure those services are "culturally competent, acceptable, and accessible to the people who need them most."

Oster also assured the audience that the CDC and health departments across the country are engaging and will continue to engage communities to hear their concerns and priorities and identify ways to work together to achieve common goals.

A Key to Ending the Epidemic

Oster wrapped up her presentation by discussing the Ending the HIV Epidemic initiative that was recently announced, which has a goal of reducing HIV transmission by 75% in the next five years and at least 90% in ten years. She noted that rapid detection of clusters is one of the four main strategies, which also include diagnosing people as early as possible after infection, treating the infection rapidly and successfully to achieve viral suppression, and protecting people at risk for HIV using potent and proven prevention techniques, including PrEP.

"I hope I have convinced you that cluster detection and response can help bring the nation closer to ending the HIV epidemic," she said in conclusion.