There is a one in five risk of  men who have sex with men (MSM) being treated for HIV passing on the infection to their male partners if they do not use condoms.

Using data from a cohort of MSM in the Netherlands, researchers used mathematical modelling (stochastic individual-based simulation model) to predict the probability that a man will infect his uninfected male partner over the course of first-line antiretroviral treatment. Three clinical situations were used: viral suppression achieved with good adherence to treatment; suppression achieved but adherence poor; suppression not achieved. Only those whose partnership was maintained throughout the duration of first-line therapy were included.

There were a few technical adjustments made based on assumptions of HIV transmission from previous studies and modelling, the number of sexual encounters per year in the partnership and efficacy of condoms in reducing transmission. Four options for condom use were considered: never using condoms; using condoms 30% of the time;

not using condoms if viral load was undetectable at the last measurement in the past six months; and always using condoms.

Men with HIV on treatment have a 22% risk of infecting their male partners if they have sex without condoms. The authors explained this finding as follows: treatment is not sufficient to suppress viral load even if the regimen is changed quickly in response to viral activity; the level of virus can rebound quickly and reach high levels before detection and change of regimen; even with viral suppression, the risk of transmission is not zero and the cumulative risk of transmission from unprotected sex becomes substantial.

In the group who used condoms 30% of the time, the risk of transmission reduced only marginally to 17%. Men who used condoms but stopped if their viral load became undetectable at the last measurement in the past six months only had a 3% risk of transmitting HIV. The risk of HIV transmission in those who always used condoms was 1% - the risk is not zero as condom efficacy is not 100%.

The frequency of monitoring also affects the risk of HIV transmission. Patients monitored frequently can switch to new drug regimens if first-line treatment fails before they expose their partner to an increased risk for a long period.

One can consider how the assumptions and stochastic modelling reflect the reality of human behaviour compared with a cohort or observational study. However, a cohort study has ethical problems of exposing an uninfected partner to HIV risk.

There are two very important messages: condoms remain the best way to reduce the risk of HIV transmission but their efficacy is not 100%; and risk of HIV transmission reduces with more frequent monitoring as this guards against risk of viral load increasing since the last measurement, i.e. decisions on condom use can be made based on a recent viral load measurement taken three months before rather than six months or longer.

This paper provides a good rationale for any practitioner who sees MSM who are HIV positive and ask about risk of transmission to their uninfected male partners. They should always use a condom regardless of viral load and get the viral load checked regularly at their HIV clinic.

Dr Richard Ma