Abstract Body

Viral sequence data provide a powerful tool for investigating closely-related HIV infections, but previous approaches have usually used only one sequence per patient, and have been unable to reconstruct the likely direction of transmission between individuals. With larger within-host genetic datasets now available, methods can be refined to incorporate inference of directionality. We have implemented such an approach in our tool phyloscanner. Here, we use phyloscanner to explore the age disparity between sources and recipients of infection amongst European men who have sex with men (MSM).

Phyloscanner was used to analyse Illumina short-read sequence data from the BEEHIVE study, sampled from 2892 Europeans living with HIV diagnosed between 1985 and 2015. The results allowed us to assign a measure of confidence to which individuals were involved in transmission pairs and in which direction transmission occurred. We selected a confidence threshold to define probable pairs and examined the distribution of transmitter and recipient ages at the time of recipient seroconversion. We also performed a regression analysis on a richer dataset where no threshold was applied but pairs were weighted by confidence level, and modelled transmitter age as a function of recipient age.

57 MSM transmission pairs were identified. In 38 (67%), the recipient was older than the transmitter, with a mean age difference of 2.16 years. However, there was marked variation in age disparity with respect to recipient age. In pairs where the recipient was under 30, the transmitter was a mean 6.11 years older. From ages 30 to 39, the mean disparity was not significantly different from zero (0.35 years). In older recipients, the transmitter was a mean 8.41 years younger. The regression model predicts that transmitter age grows by 0.28 years for every increased year of recipient age, with the cross-over to transmitters tending to be younger than recipients occurring at an age of 35.3.

Our results suggest that MSM in their thirties were more likely to be the source of infections in others of all ages. This suggests a trade-off, with MSM under 30 and above 40 both less likely to act as transmitters. In the young, this is likely due to a shorter period of potential exposure, while in the old, it could be due to reduced sexual activity or reduced unsafe sexual activity. This study demonstrates the power brought by large genetic datasets to the investigation of demographic correlates of transmission.