Tracing HIV Spread With Genetic Data in Wisconsin, a Lower-HIV State, 2014–2017

When someone is diagnosed with HIV, public health interviews (partner services) ask them to name their sexual or needle-sharing partners, which helps map transmission networks and target prevention. Genetic sequence data from routine drug resistance tests can add to this: sequences that are highly similar ("linked") form molecular clusters, and can show whether named partners are plausible transmission partners.

Such data had mostly been studied in states with more HIV. Wisconsin, with 4.6 diagnoses per 100,000 people aged 13 and older in 2016, is a lower-morbidity state. Its Division of Public Health analyzed sequences reported from 2014 to 2017.

How

• Labs must report sequences from drug resistance testing to the state, which keeps them in a secure surveillance database.
• Secure HIV-TRACE — a web tool developed by CDC with the University of California, San Diego and Temple University — compared sequences pair by pair and linked those at a genetic distance of 0.015 substitutions per site or less. Two or more linked sequences made a cluster.
• Named partners came from PartnerServicesWeb, a CDC-developed database of interview results; only pairs where both people had a sequence were compared.

What it found

Of 1,401 people with a reported sequence, 433 (30.9%) were linked to at least one other person — 703 links in 119 clusters of 2 to 20 people. Of those linked, most were male (88.5%), non-Hispanic Black (57.3%), men who have sex with men (80.8%), and aged 20–29 (50.3%).

People were mostly linked to others like themselves:

Among 123 Black men who have sex with men aged 20–29, 57.7% were linked to others aged 20–29 and 19.2% to people aged 13–19.

Interviews and genetics overlapped only partly. Of 139 named-partner pairs where both had a sequence, 47 (33.8%) were also genetically linked — plausible transmission partners.

What it means

Wisconsin's patterns — about one in three people genetically linked, mostly within the same race, risk and age group — match the national analysis, which drew mostly on higher-morbidity states. So combining sequence and interview data to find clusters works just as well in lower-morbidity states, and can reveal networks that interviews alone miss.

Genetic data cannot show who infected whom, and cluster analysis is not meant to confirm transmission; it shows population-level patterns that guide prevention.

In Wisconsin, most new diagnoses happen in outpatient clinics rather than testing sites, so people are often already in care, with resistance testing done, by the time of their interview. Health workers can then prioritize fast-growing clusters and their partners — for example, referring HIV-negative partners at high risk for preexposure prophylaxis. Interviews remain vital for finding people at high risk and undiagnosed infections, and keeping people with HIV in care.

Limits

• Clusters miss people who don't know they have HIV, aren't in care, or weren't tested for resistance or had no sequence reported.
• Elsewhere, resistance testing is less likely among older and Black people and in smaller communities; that was not checked in Wisconsin.
• The comparison with named partners covers only people who named partners.
• Risk information was imputed for the 13% missing it.

Sources

Based on Grande KM, Schumann CL, Ocfemia MCB, Vergeront JM, Wertheim JO, Oster AM, "Transmission Patterns in a Low HIV-Morbidity State — Wisconsin, 2014–2017," MMWR Morbidity and Mortality Weekly Report, volume 68, Centers for Disease Control and Prevention; a work of the United States government in the public domain.