icon-    folder.gif   Conference Reports for NATAP  
  21st Conference on Retroviruses and
Opportunistic Infections
Boston, MA March 3 - 6, 2014
Back grey_arrow_rt.gif
Novel Network Score Tied to HIV Transmission Risk and Outcomes
  CROI 2014, March 3-6, 2014, Boston
Mark Mascolini
An HIV transmission network score (TNS) deployed by researchers at the University of California, San Diego (UCSD) correlated with transmission risk behaviors and outcomes in a 648-person study [1]. The UCSD team believes the score "can be used to identify and target effective prevention interventions, like antiretroviral therapy, to those at a greater risk for HIV-1 transmission."
A number of studies have focused on defining HIV transmission clusters in defined populations, but most studies stop short of considering transmission network findings as a tool to prevent further transmission. Susan Little and UCSD colleagues addressed that issue by applying a TNS to a transmission network involving people in the San Diego area. They aimed specifically to reconstruct the local San Diego HIV-1 transmission network and to determine whether they could use the network's structure to weight HIV transmission risk, with an eye toward more efficient targeting of prevention interventions.
Multiple sites around San Diego have offered screening for acute HIV infection since1996. Researchers enrolled people with acute or early HIV infection and their contacts in the study cohort. The investigators calculated an estimated date of infection for recently infected people and assigned transmission direction by comparing that estimated date and sampling dates of putative sex partners. They inferred a partial transmission network on the basis of nucleotide genetic distances between HIV-1 pol sequences.
Then for each individual in the network the researchers calculated a TNS designed to estimate risk of HIV transmission from a newly diagnosed person to a new partner. TNS, the UCSD team explained, is "the proportion of nodes [individuals] in the network with degrees less than the newly enrolled individual."* Finally, using the network at the time of enrollment for each individual, the researchers figured the correlation between baseline TNS and putative subsequent transmissions in the year after enrollment.
Analysis of HIV-1 pol sequences inferred a transmission network of 648 people, 478 of them (74%) with recent infection and 170 of their contacts. Putative HIV transmission correlated with higher plasma viral load (5.2 log10 copies/mL with putative transmission versus 4.7 without, P < 0.01, that is, about 158,000 copies versus 50,000 copies), with more sex partners (3.0 versus 1.5, P = 0.03), and with TNS above 0.75 versus lower (44.8% versus 15.6%, P < 0.01).
In multivariate analysis, higher viral load and more sex partners approximately doubled odds of HIV transmission (adjusted odds ratios 2.0, P < 0.01, and 1.8, P = 0.03), while a TNS above 0.75 quadrupled transmission odds (adjusted odds ratio 4.0, P < 0.01).
The UCSD investigators concluded that TNS correlates with transmission risk behavior and outcomes and proposed that it can be used to identify high-risk individuals who will benefit from interventions such as antiretroviral therapy. They noted that universal testing and treatment "remain the goal of an effective HIV prevention strategy." But without universal test-and-treat, they added, targeted prevention of people with a high risk of transmission (such as those with a high TNS) "may offer a more efficient use of prevention resources and provide a greater reduction in network incidence."
*TNS(d/N) = probability (degree of a node in N < d), with the probability computed using the best-fitting parametric density for the network N. A node is an individual in the network.
1. Little SJ, Pond SLK, Anderson CM, et al. Using HIV networks to inform real time prevention interventions. CROI 2014. Conference on Retroviruses and Opportunistic Infections. March 3-6, 2014. Boston. Abstract 206.