Background: Knowing the fraction of transmissions attributable to recent HIV infections is essential for the success of Treatment-as-Prevention. This is because recently infected patients are often unaware of their HIV status and hence remain untreated and highly infectious.
Methods: A maximum-likelihood phylogeny was constructed from 121,306 HIV-1 pol sequences (21,471 sequences from 11,567 Swiss HIV Cohort Study (SHCS) participants and 99,835 sequences from the Los Alamos database). Swiss transmission clusters were identified using different combinations of intra-cluster genetic distance (1%, 1.5%, 2%, 2.5%) and bootstrap (50% to 100% by increment of 2%) thresholds, in order to determine the effect of those criteria. Seroconversion dates were estimated based on immunological markers, dates of HIV positive/negative tests, clinical symptoms and ambiguous nucleotides. Transmission clusters were classified as recent or chronic transmission based on the maximal time interval between the seroconversion dates of the cluster members. Logistic regression with adjustment for age, sex, risk group, HIV subtype, baseline RNA/CD4, time-to-ART and Chronic phase RNA viral load integral, was applied to identify, among transmitters, the risk factors associated with having transmitted in the recent or chronic phase.
Results: Depending on the implemented phylogenetic threshold criteria, we identified between 50 to 271 transmission clusters with known seroconversion dates for all members. These clusters showed that:
1. The median fraction of transmission during recent infection was 42.2% (range 37%-56%) when recent infection was defined as the first year of the infection and 32% (range 28%-43%) for a six months definition.
2. Stricter criteria for defining transmission cluster (higher bootstrap thresholds) were strongly associated with higher fractions of recent phase transmission, Figure 1, (Spearman`s rho 0.95, P <0.001).
3. The total viral load in the chronic phase (measured as the Chronic phase RNA viral load integral) was negatively associated with recent as opposed to chronic phase transmission, OR 0.43 (0.24-0.77). This effect was even stronger in the adjusted model OR 0.25 (0.1-0.67).
Conclusions: Our data points to a high fraction of transmission during recent HIV infections in the SHCS. Moreover, we show that the total viral load in the chronic phase is a strong determinant of the transmission phase. These results are important to the pertinent debate on Treatment-as-Prevention as an “Endgame” strategy.