Estimated HIV risk per exposure

Healthcare professionals are constantly asked to give precise estimates for the likelihood of infection. Categories like 'high risk' and 'low risk' can be perceived as evasive, unhelpful and frustrating for people in search of precise, numeric answers.

This page therefore lists a number of estimates (or guesstimates) of the likelihood of infection per exposure for specific activities and incidents. The figures are drawn from the most useful cohort studies and meta-analyses (which pool the data from as many studies as possible). Nonetheless, close examination of these studies often reveals methodological problems and limitations in the data. The studies do not all follow the same methodologies, making comparisons between figures difficult.

In this section, risk figures are expressed as per exposure estimates. Per exposure means the risk of infection for each individual incident during which HIV transmission could have occurred; e.g. for each separate act of intercourse. This is a more precise measure of risk than ‘per sexual partner’, which does not take into account the number of times sex happens and the fact that sexual partners may or may not have HIV.

The per-exposure measure of risk may cause activities to seem less risky. You might be surprised at the apparently low figures that constitute 'high risk' activities. Being told that there is, for example, a one in 200 chance of infection could, conceivably, lead someone to think, “Only one in 200. Well, that’s not too bad”.

This figure does not take into account the fact that people do certain things (e.g. have sex) a lot more often than they do other things (e.g. prick themselves with an infected needle). This 'one in 200' figure means that the person would only have to have sex with the source partner 140 times for it to become more likely than not that they will catch HIV.

These figures assume an HIV-positive partner. For a partner of unknown status, the risk is affected by the prevalence of HIV in the relevant community – i.e. the chance that the partner does in fact have HIV.

A key factor determining the risk of transmission is the amount of virus in body fluids, which is known as viral load. When a person is living with HIV and is on effective treatment, the amount of HIV in their body fluids falls drastically, to a level that is described as ‘undetectable’ viral load. At this stage, HIV cannot be passed on sexually.

On the other hand, a very high viral load increases the risk of HIV transmission. Each log₁₀ increase in viral load is estimated to increase transmission risk threefold (relative risk 2.89). Very recent (acute) HIV infection, a time when viral load is exceptionally high, increases the risk sevenfold (relative risk 7.25). Similarly, advanced HIV disease increases the risk sixfold (relative risk 5.81).

Another factor that would increase the risk of sexual transmission is the HIV-negative person having an STI (genital ulcer disease). This is thought to more than double the chances of acquiring HIV (relative risk 2.65).

The risk can be lowered by prevention measures. When looking at the table below, it is worth bearing in mind that the available figures are not always directly comparable. Some figures are from carefully managed clinical trials, whereas others are from real-world conditions. Some facts can be biologically verified (such as an undetectable viral load) whereas others rely on how people report their behaviour to researchers (such as condom use). For example, if people who occasionally use condoms tell researchers that they always use them, the results for ‘consistent condom users’ included below will include data from people who are actually inconsistent users. These individuals are more likely to acquire HIV and condom effectiveness will be underestimated.