Efficacy

Because large-scale prospective, placebo-controlled trials have not been carried out for PEP, a definitive answer regarding its effectiveness cannot be given, but experience from several PEP programmes has been recorded in a number of studies.

A study in San Francisco¹ investigated the safety and acceptability of PEP after possible sexual exposure to HIV or exposure involving injecting drug use (IDU). Individuals who had experienced exposure within the previous 72 hours were offered one month's treatment with Combivir. Other drugs, ddI (didanosine, Videx) and d4T (stavudine, Zerit), were offered in cases where there was a risk of exposure to AZT- or 3TC-resistant virus.

In total 401 requests for PEP were received between December 1997 and March 1999: 91% of participants were men and the median age was 32 years.

The vast majority (93.5%) of the participants sought PEP because of sexual exposure and only 2% reported sharing of IDU equipment. Receptive anal intercourse was reported in 40% of the exposures.

Certainty that the source partner was HIV positive was expressed by 174 (43%) of participants. When exposure was sexual, it usually represented a lapse in safe sex practices rather than habitual high-risk behaviour.

Complete adherence to medication (in the four days before a clinic visit) was reported by 84 to 78% of participants, despite high levels of side-effects, including nausea, fatigue, headache and diarrhoea.

Completion rates were high at 78%. This was most likely due to the provision of one-to-one medication adherence counselling, and to dispensing only a limited supply of medications at each visit - which required individuals to make regular contact with staff.

The study was not powered to evaluate efficacy, though no individuals were observed to develop antibodies to HIV at six months after exposure.

A subsequent study in San Francisco in 2005² evaluated 702 subjects who had taken PEP after sexual exposure (including some of the subjects enrolled in the study discussed above). The evaluation took place twelve weeks after exposure. Three people seroconverted despite reporting no further sexual exposures after starting PEP, and so are probably evidence of PEP failure. In addition, four people became HIV-positive despite PEP but there were either additional exposures to HIV, or HIV was found in blood specimens given at the start of PEP (meaning subjects were already infected with HIV and did not know it).

Of course, given that the chance of contracting HIV from a single occasion of receptive anal intercourse is in the region of one chance in 30 to one in 250, many of the men taking PEP would not have seroconverted anyway. Because of this, among 702 men one would only expect about six infections if they only had one risky encounter each. But because only three became infected who had definitely only had one encounter, the study implies that PEP prevents at least one in two infections.

The average time delay between the sexual exposure and taking PEP was 32.5 hours, but among men who became HIV-positive it was 45.5 hours (nearly two days) lending support to the view that initiating PEP within a day of exposure ensures it has the best chance of working.

Researcher Michelle Roland said that the four men who went on to have unsafe sex soon after taking PEP were a reminder that it could not be a stand-alone prevention method: “It provides an important HIV prevention opportunity when exposed individuals may be especially receptive to assistance with reducing their HIV risk...but is only one part of that prevention activity. Sexual exposures are usually not isolated, and helping people stay HIV negative requires response to both the presenting exposure and attempts to reduce subsequent exposures.”

One of the most conclusive studies of PEP was conducted in Brazil and reported at the 2002 Conference on Retroviruses and Opportunistic Infections (CROI).³ Here, 202 HIV-negative gay men were enrolled and followed for an average of two years. At the time of enrolment, 57% of the group reported “high-risk behaviour”. PEP, consisting of four weeks of AZT/3TC, was used 100 times by 73 (36%) of the participants, 91% of whom completed the course. Most men took it just once or twice during the two years of the trial, but one man took it nine times.

There were 11 seroconversions among the group; however, only one occurred in someone who had taken PEP. Analysis of the strain of HIV that infected him despite using PEP showed that it had the M184V mutation (associated with resistance to 3TC).

The researchers calculated that PEP, on the occasions it was taken, reduced the seroconversion rate by 83%, from 4.1 cases per 100 patients a year to 0.7 cases when men did take PEP.

However the study also found that there was almost no difference in the infection rate observed in the study population as a whole (2.9 cases per 100 patients a year) and what would have been expected if PEP had not been available (3.1 infections per 100 patients a year). This was because PEP was not taken nearly often enough to make a difference to HIV incidence in the group as a whole. Although quite a high proportion of the study population took PEP, they were bad at evaluating risk and did not always take it when needed.

The reasons PEP was not taken were either that the men mistakenly assumed their partner was faithful and HIV-negative, or that they caught HIV through routes they considered low risk, such as oral sex.

This seems to bear out findings from the UK about gay men’s estimation of risk. The 2004 Gay Men’s Sex Survey, Risk and Reflexion,⁴ found that gay men tend to see sexual risk in terms of risky actions rather than risky partners. Their focus was on condoms or different sexual acts rather than the HIV status of their partner. Sigma Research, which conducted the research, comments that this gives an insight into where men go wrong in evaluating risk: “The problems seem to be more with misreading the presence of HIV infection (exposure) than underestimating the potential for transmission when it is present.”

The Brazilian study highlighted one of the central dilemmas of PEP. Unlike other interventions, its availability may be crucial to individuals who need it but, as an emergency measure only used by a minority of gay men, it is unlikely to make much difference to HIV incidence on a population level. This has led to controversy about exactly how much it should be promoted as a prevention method – see more below.

However because people seeking PEP are highly motivated to avoid infection, this may be a time when education and counselling will have significant influence.

Several other interesting findings came from the Brazilian study. One was that “high-risk” behaviour declined during the study from 56 to 40%. Although it has been suggested from other studies that people reporting declines in risky behaviour may sometimes not be reporting the whole story due, possibly, to guilt at being a ‘study failure’, this finding may help allay fears that the widespread availability of PEP for sexual exposure will cause an increase in condomless intercourse.

Furthermore, 92% of the time, PEP was taken appropriately, i.e. after an exposure that researchers considered high risk. This seems to offer a degree of assurance that PEP will not be abused by the ‘worried well’ concerned by very low-risk incidents.

Finally, the study issued people with ‘starter packs’ – ready-wrapped doses of AZT/3TC with instructions that participants should take them immediately following exposure. Other PEP schemes have relied on people reporting to A&E departments after sexual exposure. Since one of the key components to successful PEP is prompt treatment (ideally within 24 hours) this may explain some of the difference between the reported effectiveness rate of 83% in the Brazil study and the 53% rate in the cost-effectiveness study detailed in the following section.

Even failed post-exposure prophylaxis may have some effect. A study in 2008 suggested that unsuccessful PEP may still result in weaker HIV infection and lower viral load.⁵ Although this is only a case report, it does follow reports of similarly suppressed HIV viral loads in monkeys who were infected with HIV in studies of pre‑exposure prophylaxis.

The case report involved a patient who received PEP with Truvada a fixed-dose combination pill containing FTC and tenofovir. Although this treatment failed to prevent HIV infection, the patient did have a well-preserved immune function and a lower viral load than would be expected. He was therefore much less infectious than the average patient during acute infection.

The 38-year-old New York gay man first came to a clinic on 26 September 2006 reporting that he had had unprotected receptive anal sex with multiple partners during the previous 48 hours. He was treated with Truvada but during the period on this treatment, on 24 October, he reported more episodes of risky sex, and his course of PEP was therefore extended. He stopped taking it on 7 November and tested HIV-negative on that date.

He reported a third episode of risky sex on 28 November and was restarted on Truvada. On 18 December, three weeks later, he tested HIV-positive. He was adamant that he had had no risky sex during the period when he was not taking post-exposure prophylaxis, which he finally stopped taking after he tested HIV positive, on 29 December.

His first two viral load tests were performed on 22 December (when he was still receiving treatment with Truvada) and on 9 January 2007. His viral load was very low, with 213 and 647 copies/ml in these two tests respectively. His viral load increased after this but never exceeded 3500. His CD4 count was a very high 1800 cells/mm³ on 22 December, just after his first positive HIV antibody test, and then fell to about 750. At no time did he have the high viral load and low CD4 count typical of acute HIV infection, and he had no HIV seroconversion symptoms.

The patient's antibody response developed much more slowly than normal. Some basic tests were performed on his HLA genes, which determine susceptibility to HIV infection, but he had no genetic mutations associated with a lower viral load or less virulent course of HIV infection.

One theory of how HIV causes AIDS is that the initial destruction of CD4 cells and immune hyperactivation in the gut, from which the body never completely recovers even under HIV therapy, eventually depletes the immune system. However tests suggested that this man’s gut immune system was better preserved than in other cases of acute infection. It is possible that disease progression may therefore be slower.