Key studies

In May 2009, Attia and colleagues published a systematic review and meta-analysis examining all known prospective studies published or presented between January 1996 and February 2009 on the risk of HIV transmission through unprotected sexual intercourse according to viral load.¹ All were in heterosexual couples. There have been no prospective studies examining the relationship between viral load and risk of transmission in sex between men, nor during anal sex, which is also practised by a significant minority of heterosexual couples, and which is often not reported, particularly in Africa where the practice is especially considered to be taboo.²

Of the ten studies that included HIV-positive individuals not receiving antiretroviral therapy the overall HIV transmission rate over 9998 person-years of follow-up was found to be 5.64 per 100 person-years. They calculated that amongst people with a viral load below 400 copies/ml the transmission rate was 0.16 per 100 person-years. The transmission rate increased with increasing viral load to 9.03 per 100 person years amongst individuals with a viral load of at least 50,000 copies/ml.

This means that in 1000 people with viral load below 400 copies/ml, at least one transmission could be expected to occur in the course of a year. In contrast, among 1000 people with viral load above 50,000 copies/ml, at least 90 transmissions could be expected to occur.

The review also included new information concerning a prospective observational study involving 393 monogamous heterosexual couples in Spain to determine HIV transmission risks, originally published in 2005. It revealed that there had been one case of sexual transmission when a HIV-positive partner, who was not on treatment, had a viral load of 362 copies/ml.³

A study from Uganda, often considered to be the benchmark study confirming that viral load measured in the blood is the most important factor in determining whether or not HIV is transmitted following sexual exposure, previously found that no HIV transmission was observed over a 30-month period in the 51 couples where the HIV-positive partner had consistent viral load measurements in the blood below 1500 copies/ml.⁴ Two further studies have since reported HIV transmission when viral load was below 1500 copies/ml (again only in untreated individuals): at viral loads of 600 copies/ml⁵ and 1497 copies/ml.⁶

Table 1 summarises all of the prospective studies published or presented so far on HIV transmission according to viral load.

Table 2 summarises HIV transmission rates per 100 person-years according to viral load, where the HIV-positive partner was not on treatment.