How new understandings of AIDS pathogenesis and improvements in antiviral treatments will impact on HIV primary prevention.

This is a fully updated version of an article originally published in Acting On AIDS (ed. Oppenheimer & Reckitt), Serpent's Tail, 1997. Reproduced with permission from Edward King's AIDS Pages

By Edward King

Since January 1995, developments in HIV basic science and clinical research have come thick and fast. Important advances have been made in understanding the natural history of HIV within the infected person's body and in developing anti-HIV treatments whose effectiveness is orders of magnitude greater than those previously available. Already the medical arena is being transformed, as many clinicians feel increasingly confident that their interventions can have a dramatic impact on the quality and quantity of life of people with HIV.

However, there has been little consideration of the ways in which the new virology will affect HIV prevention policies and practice. This paper argues that it will inevitably result in unprecedented upheavals. Even if the effectiveness of today's therapies is still insufficient to bring to fruition the more optimistic aims, such as totally halting disease progression or even curing infected people, the fact that these possibilities are now openly discussed in medical circles itself has implications for HIV prevention work. Until very recently in the UK, treatment issues were largely ignored or overlooked by the AIDS voluntary sector (King, 1997a), and even today few are aware of their significance other than for HIV-infected people. Now more than ever it is essential that HIV prevention workers understand the significance of recent medical advances and prepare for the impact on their own work.


THE NEW VIROLOGY

New insights into the relationship between HIV and the human immune system have been central to the latest developments. Throughout the first decade of the epidemic it was generally believed that shortly after infection with HIV the body mounts a strong immune response that drives the virus into latency, and that the virus remains relatively inactive throughout the asymptomatic period. Only as the disease advanced was HIV thought to be re-activated by some unknown trigger and to start replicating at high levels. In January 1995 two American research teams produced evidence that this was not the case (Ho , Wei). Their studies relied on two new tools ­ viral load tests and protease inhibitors ­ and an understanding of the process of drug resistance.

There are several different viral load tests ­ or quantitative HIV RNA assays, to give them their precise technical name ­ but they all essentially measure the same thing. Using a variety of amplification techniques, they ëcount' the number of pieces of HIV RNA in a sample, usually blood plasma. RNA is the form in which HIV particles (and other retroviruses) normally carry their genetic material. When HIV infects a cell, it converts its RNA to DNA (the same form as the genetic code of human cells) and merges with the host cell's genes. HIV RNA assays therefore measure the number of HIV particles that are present in the sample, as opposed to the number of HIV-infected cells. Other assays that detect HIV DNA - in other words, which count the number of HIV-infected cells - are available to research laboratories but are not yet used in routine clinical practice (King, 1997b).

Protease inhibitors are the newest class of anti-HIV drugs to be developed (Moyle, 1996). Three protease inhibitors have already advanced to the clinic ­ saquinavir (Invirase) manufactured by Roche, Abbott's ritonavir (Norvir), Merck's indinavir (Crixivan) and Agouron's nelfinavir (Viracept). As a class, protease inhibitors have substantially greater anti-HIV effects than earlier drugs such as the nucleoside analogue family that includes AZT, 3TC, ddI and ddC.

However, when exposed to ritonavir or indinavir as single drug treatments (monotherapy), HIV strains that are resistant to the effect of the drugs rapidly evolve. Resistance is thought to occur primarily through a process of Darwinian selection ­ ësurvival of the fittest'. HIV replication is a relatively error-prone process, so each new HIV particle produced from an infected cell contains subtle genetic variations from the parent virus. Some of these naturally occurring mutants may, by chance, be inherently less susceptible (or resistant) to the effects of anti-HIV drugs than others. When HIV reproduces within the body of someone taking a drug, any susceptible viruses that are produced are likely to be eradicated by the drug, but resistant viruses are more likely to survive to infect new cells and themselves reproduce. Over time, this ëselective pressure' may mean that an initially drug-susceptible virus population evolves into a predominantly drug-resistant population (Moyle, 1995).

The American research teams examined the short-term impact of a protease inhibitor on levels of HIV in the blood of infected people. They found that plasma viral load rapidly declined within hours of starting treatment. Protease inhibitors only block the production of new HIV particles from HIV-infected cells; they do not kill HIV particles that already exist in the circulation. The fact that the number of virus particles in the blood fell so rapidly once the production of new viruses from infected cells was stopped by the drug thus indicated that the ëlife expectancy' of a free-floating virus particle in the blood must itself be very short.

Substantial numbers of virus particles can be measured even in the blood of healthy, untreated, symptom-free HIV-positive people. Because the life-span of such particles is short, their presence proves that high levels of HIV reproduction must be taking place throughout the period in which HIV was previously thought to be primarily latent.

During this asymptomatic period, though, people with HIV tend to have relatively stable CD4 cell counts. CD4 cells, or T-helper cells, are the primary immune cell infected by HIV and destroyed either by the virus or by other immune cells that recognise infected CD4 cells. The only way that the CD4 count could remain stable even though a large number of HIV particles is being produced every day is if the immune system is producing new CD4 cells at about the same rate as they are destroyed.

Thus, a picture emerges of the virus and the immune system engaged in an ongoing, high-level battle throughout the asymptomatic period. Every day billions of HIV particles are produced, and just as quickly as these viruses destroy CD4 cells, the immune system valiantly produces more CD4 cells to replace them. The apparent stability of the CD4 count and viral load over periods of many months belies the true level of viral and immune activity; it is like trench warfare viewed from the air, in which the battle-lines barely move but both sides constantly suffer massive casualties and must marshal a steady stream of reinforcements.

Within a few months of infection, an individual's viral load has reached this state of equilibrium with the immune system; the level of the viral load at this point is called the ëset-point', and it is thought to remain relatively stable until significant disease progression occurs. The viral set-point has been shown to be strongly correlated with the rate at which an individual's CD4 cell count subsequently declines and with their medium- to long-term risk of developing AIDS and dying. In a study of stored blood samples from over 1600 gay men in the Multicenter AIDS Cohort Study in the USA, people with the lowest viral load (below 3,000 copies/ml) soon after infection took an average of over ten years to develop AIDS, while those with the highest viral load (over 30,000 copies/ml) took an average of only 2.8 years to develop AIDS (Mellor s).


THERAPEUTIC ADVANCES

Since a low viral load is associated with a good prognosis among untreated people, many researchers argue that reducing a high viral load with anti-HIV drugs (either before the set-point is established or at any subsequent stage in the course of infection) should be expected to improve prognosis. This concept has not been definitively proven, but is supported by the recent results of several trials studying the effects of drug combinations on HIV-related symptoms and life expectancy. These have consistently indicated that the regimen that leads to the greatest reduction in viral load within the first one to two months of therapy is also the most effective at reducing disease progression and prolonging life in the long-term.

With two or especially with three-drug combinations, dramatic falls in viral load can be achieved. In some studies the majority of treated people have had their viral load reduced to levels that are undetectable with even the most advanced viral load assays. This does not necessarily mean that there is no virus present at all ­ just that virus levels are below the limit of detection of the test used. For example, the most widely used viral load test, the Amplicor HIV Monitor made by Roche, currently has a lower limit of detection of 400 copies/ml. Samples whose viral load is anywhere between 399 and zero copies/ml will all be below the limit of detection, or 'undetectable'.

Even if treatments were successful in reducing the level of HIV RNA in a sample of blood to zero, this would not indicate that the individual was no longer infected (or infectious). First, as noted above the RNA assay only detects HIV particles, rather than HIV-infected cells which can be detected by testing for HIV DNA. All the studies conducted to date have found that HIV DNA can still be detected in samples of blood or body tissues, even among people who have had undetectable RNA in their blood for many months. Some argue that these may be defective viral material that is not capable of producing new viruses. However, the most prudent interpretation is that HIV-infected cells persist, retaining the potential to resume production of HIV particles if drug therapy is stopped or its effectiveness diminishes, and to be passed on to others through unsafe behaviours. Secondly, it is possible that HIV replication is not being suppressed as effectively in parts of the body to which the drugs may not penetrate as well, such as the brain, spinal cord and testes, and will re-emerge if treatment is stopped or multi-drug resistant HIV strains develop.

A rebound in plasma viral load has indeed been observed in every case to date in which someone whose viral load had been reduced below the level of detection has elected to stop treatment. In one case reported at the Eleventh International Conference on AIDS in July 1996, a man who had had no detectable HIV in his blood or lymph nodes for 18 months decided to stop treatment ­ and the virus promptly reappeared (Saget). Nevertheless, the more optimistic researchers argue that treatment with the most powerful drug regimens available might theoretically eliminate HIV from the body altogether. They reason that a regimen that totally suppressed viral replication throughout the body would prevent the production of any new virus particles, and that cells that are already infected with HIV would eventually reach the end of their natural lifespan and die - taking the virus with them - to be replaced by uninfected cells. However, it is estimated that it would probably take three years or more for some of the longer-lived cell types infected by HIV, such as macrophages, to die and be replaced.

Experiments designed specifically to test the ability of three and four drug combinations to cure newly infected people are underway in New York. The researchers argue that people who have been infected for only a few months are the best candidates in whom to try to eradicate the virus, for three reasons: because HIV may not have had a chance to spread widely throughout their body tissues; because relatively few HIV life-cycles will have taken place, limiting the number of potentially drug-resistant mutants that are likely to be present; and because these individuals are unlikely yet to have suffered irreversible damage to their immune systems.

It will take many years for these studies to reach a conclusion. However, even if early aggressive treatment does not cure newly infected people, many researchers argue that it may still dramatically alter the long-term course of infection. As noted earlier, among untreated people, those with a low viral load a few months after infection are much less likely to develop AIDS during the next ten years than those with a higher viral load. Researchers hope that reducing the set-point of viral load through the prompt use of anti-HIV drugs at or soon after seroconversion will also improve long-term prognosis, although this remains unproven.


IMPLICATIONS FOR PREVENTION

Treatment information providers have responded quickly to the urgent need to inform people with HIV, their doctors and those responsible for funding treatment and care services about these advances. Yet the implications extend far more widely than this. Three major issues for HIV prevention workers can already be identified. First, how will improvements in the treatment of HIV affect the willingness of people who may have been at risk of infection to take the HIV antibody test? Secondly, what is the significance of evidence that prompt treatment soon after exposure to HIV can reduce the chance of becoming infected? Thirdly, what does the new virology have to tell us about the infectiousness of people with HIV, and ways in which it may be reduced by drug therapy?


TESTING

The lack of anti-viral treatment options that might delay the onset of AIDS has been a strong disincentive against taking the HIV antibody test. Many individuals and HIV advisors have argued that there is little point in finding out whether one is HIV-positive if there is nothing that can be done to delay the development of symptoms. In fact, there are medical interventions that can benefit asymptomatic people with HIV, such as the use of antibiotics to prevent PCP pneumonia among people with CD4 counts below 200, but in the UK at least these have never received much attention as an argument in favour of testing. Just as doctors report that a growing proportion of people already diagnosed with HIV is choosing to take anti-HIV drugs, so awareness of the availability of combination therapy regimens that can delay AIDS and prolong life when started before symptoms occur is likely to encourage a growing proportion of previously untested people to find out their status. The ability of viral load testing to predict an individual's likelihood of disease progression (especially when used in combination with the CD4 count) may provide a further incentive for testing among symptom-free people, who in the past might have concluded that testing HIV-positive would introduce more doubts and uncertainties than if they remained untested.

Many clinics appear to be adopted a informal policy to try to identify recently infected people and offer them aggressive anti-HIV therapy during the period known as ëprimary' or ëacute' infection. As discussed earlier, the set-point at which an individual's viral load settles three to six months after infection is a good predictor of their long-term prognosis, with a low set-point clearly associated with a reduced risk of disease progression. Several studies suggest that using combination therapy in recently infected people can reduce viral load to undetectable levels in a high proportion of treated people, raising the hope that this will also dramatically improve their prognosis. Although many questions remain unanswered for instance, is a viral load that has been lowered with drugs as good a prognostic marker as a viral load that is low without treatment? Will viral load rebound if treatment is stopped? ­ already some American AIDS groups are initiating campaigns to raise awareness of the symptoms of HIV primary infection and encourage affected people to access medical services.

People who have previously tested HIV-negative may start to take repeat tests every few months, so that if they do become infected they are diagnosed promptly. A less frequent form of repeat testing is already commonplace among American gay men, not least on account of the requirement that federally-funded prevention campaigns include a testing component (Patton ), but on the whole community educators have insisted that testing is significant, in Simon Watney's words, "as a means of access to treatment and care, rather than as a form of primary HIV prevention" (Watney , 1994). During the 1990s, this separation of testing and prevention was challenged by the recognition of the complex strategies employed by many gay men to reduce their risk of HIV infection, such as the ënegotiated safety' approach in which men in relationships use the test to establish that they share the same serostatus, then only use condoms during sex outside the relationship. This trend towards the wider acceptance of HIV antibody testing as a useful component of HIV prevention strategies can only accelerate if the test is recognised as the means of access to new biomedical forms of HIV risk reduction such as post-exposure prophylaxis, discussed below.

This pressure towards early diagnosis may require the introduction of new tests in the clinic. The one-to-three month ëwindow period' between infection and the production of HIV antibodies means that conventional HIV antibody tests are not a useful tool for definitively diagnosing recent HIV infection. Alternative approaches include the use of direct viral assays that measure HIV RNA (viral load tests) or the presence of p24, an HIV protein that can be detected in the blood of many infected people.


POST-EXPOSURE PROPHYLAXIS

In recent months there has been renewed interest in the use of anti-HIV drugs not to treat established HIV infection, but to try to prevent the virus from gaining a foot-hold in the body of, for example, a health-care worker exposed to HIV through an accident such as a needle-stick injury. While debate on this subject predominantly focuses on the medical profession, post-exposure prophylaxis (PEP) may also be important for people exposed to HIV through sexual or drug-using behaviours.

Post-exposure prophylaxis has previously been controversial because of the limited evidence that it works. The average risk of infection from an occupational injury involving HIV-infected needles or other sharp instruments has been estimated at one in 300, or 0.3% (Tokars ), making it very difficult to show whether the prompt use of anti-HIV drugs can reduce this very small risk still further, and the only attempt at a controlled trial failed to enrol enough participants (LaFon). Nevertheless, in 1992 40% of health-care workers who suffered occupational exposure to HIV chose to use PEP (Tokars ).

In 1995 the US Centers for Disease Control and Prevention published the results from a retrospective case-control study analysing the medical records of health-care workers in the USA and Europe who reported accidental exposure to HIV-infected blood through, for instance, needlestick injuries between 1988 and 1994. The study concluded that the proportion of workers who became infected with HIV as a result of an occupational accident was reduced by 79% among those who took AZT after the exposure, compared with those who declined AZT (CDC, 1995). Other factors that significantly increased the likelihood of seroconversion in addition to not receiving AZT were exposure to a large quantity of blood, and exposure to blood from someone in the advanced stages of AIDS.

The precise mechanism by which PEP with reverse transcriptase inhibitors such as AZT reduce the risk of infection is not entirely clear. These drugs do not prevent HIV from entering cells, but they do make it harder for HIV to integrate its genetic material with that of the cell and in effect turn the cell into a factory for producing more virus particles. A certain minimum amount of virus is probably needed in the body if it is to establish a permanent infection. Treatment with reverse transcriptase inhibitors, if started soon enough after the initial exposure to HIV, may help to prevent some cells from becoming infected with HIV and thus limit the amount of HIV that is produced in the body, keeping the amount of virus below the level that can lead to proper infection. Protease inhibitors make it harder for cells that have already been infected with HIV to produce new virus particles that could go on to infect further cells, again reducing overall levels of HIV in the body. However, AZT is the only drug out of the six reverse transcriptase inhibitors and four protease inhibitors approved in the USA to have been tested as PEP in either humans or animals.

However, multi-drug combinations have been demonstrated to be more effective than AZT monotherapy in treating people with HIV. Thus, when the CDC issued new guidelines for PEP after occupational risk exposures, it suggested that combinations should also be considered in the hope of increasing the effectiveness of the treatment.

The CDC now advocates four weeks' treatment using regimens of varying intensity, depending on the severity of the exposure. For example, towards the lower end of the spectrum of risk, workers who suffer exposure of a mucous membrane to a potentially infectious fluid such as semen or vaginal secretions (in which cases the average risk of infection is estimated at 0.1% (Gerberding) should be offered AZT perhaps with the addition of 3TC. In more serious exposures, such as those involving contact between mucous membranes and blood, or when the skin is broken by a bloodied instrument, or when the blood may be suspected to contain high levels of HIV (as in cases involving newly-infected people or people with advanced AIDS), treatment with the triple combination of AZT, 3TC and indinavir may be recommended. The guidelines note that in cases where the HIV status of the 'source patient' is unknown, "initiating PEP should be decided on a case-by-case basis, based on the exposure risk and the likelihood of HIV infection in known or possible source patients" (CDC, 1996).

At a session discussing PEP during the Eleventh International Conference on AIDS in Vancouver in July 1996, physicians voiced their sense of unease at limiting PEP to health-care workers. As Katz and Gerberding point out, "The probability of HIV infection due to puncture by a contaminated needle is similar to that estimated for a single episode of unprotected receptive anal or vaginal intercourse with an infected partner or for a single episode of injection-drug use with HIV-contaminated equipment". If PEP works, ethically it should be made available to individuals who are placed at significant risk of infection through other exposure routes.


TABLE: ESTIMATED PROBABILITIES OF HIV TRANSMISSION (Katz & Gerberding)

Unprotected receptive anal intercourse	0.008 to 0.032
Unprotected receptive vaginal intercourse	0.0005 to 0.0015
Unprotected insertive vaginal intercourse	0.0003 to 0.0009
Use of HIV-infected drug injecting equipment	0.0067
Puncture by an HIV-infected needle	0.0032

Note: average figures must be interpreted cautiously as many factors may increase or decrease infectivity and/or susceptibility to infection


Use of PEP in non-occupational settings has only previously been considered for survivors of sexual assault. The American Society of Law, Medicine and Ethics convened an interdisciplinary Working Group on HIV Testing, Counselling, and Prophylaxis After Sexual Assault, which published its recommendations in 1994. At that time the efficacy of PEP was less clearly established, but nevertheless the group concluded that survivors should be provided with information about the availability of PEP to enable them to decide whether or not to use it, "based on a risk assessment of the exposure. The risk assessment should consider available information on the serostatus of the assailant, the type of exposure (anal, vaginal, or oral penetration and ejaculation), the nature of the physical injuries, and the number of assaults", as well as the potential for drug side-effects (Gostin) .

When viewed alongside the growing acceptance of treatment during or soon after seroconversion, the case for offering PEP to anyone recently at significant risk of infection becomes quite compelling. It will certainly be difficult to define the boundary between cases of sexual risk which are high enough to justify offering PEP, and those in which the risk of infection is sufficiently low that the financial cost of PEP and the risk of drug side-effects is felt to be unjustifiable. For instance, the CDC guidelines indicate that combination therapy may be reasonable for health-care workers who experience mucosal exposure to semen even where there are only grounds for suspicion, rather than certainty, that the source patient is HIV-positive. How does this differ from the situation of any gay man who gets fucked without a condom in a large city in the UK?

When used, PEP should be initiated promptly, since animal research suggests that PEP may be ineffective if started later than 24 to 36 hours after exposure. A recent discussion of PEP for non-occupational risks recommends against initating treatment more than 72 hours after the exposure (Katz & Gerberding), although the CDC guidelines argue that starting even one to two weeks post-exposure may be justified in cases of the highest risk. The protocol used at San Francisco General Hospital notes that "after an exposure, most health-care workers are upset and find that decisions about treatment are very hard to make. We recommend that the exposed person start therapy. Therapy can be stopped later, after the exposed person has had a chance to talk with their clinician and loved ones. Once the immediate crisis has passed, it is usually easier to make the best decision." (San Francisco General Hospital Epi-Center). If PEP is ever to become a practical option for non-occupational exposure, a new system of 'rapid response' clinic services may be required to provide prompt access to treatment.

A universal policy of prescribing PEP for people who have experienced any significant HIV risk exposure could never be cost-effective, even though at around £750 the cost of a month's triple combination therapy PEP for a single individual seems to compares extremely favourably with the likely life-time costs of treating the same individual should he or she become infected with HIV. The cost-effectiveness of an intervention such as PEP can only be meaningfully calculated in terms of the amount of money that would need to be spent to prevent a single infection. On average, no more than about one out of every three hundred people who have a single episode of unprotected receptive anal sex with an HIV-positive person becomes infected as a result (Katz & Gerberding). So if all 300 came forward for PEP after their risk exposure, 299 would be treated 'unnecessarily', because they would not have become infected regardless of whether or not they received PEP. If doctors have to treat 300 people in order to prevent the one single infection, the cost of preventing that infection would be three hundred times £750, which makes £225,000. In blunt financial terms, this no longer compares so favourably with the life-time costs of medical care.

The cost-effectiveness of PEP could be improved by using fewer or cheaper drugs; for example, if only two nucleoside analogue drugs were used (or if an additional protease inhibitor was reserved only for specific cases of the greatest risk) the cost per course of PEP would be approximately halved. Moreover, PEP would also be more cost-effective if it were delivered only to people whose circumstances meant that they were most at risk of becoming infected (effectively reducing the proportion of recipients who are being treated 'unnecessarily'). Possible criteria for prioritisation might include limiting PEP to cases in which people had a risk encounter with someone who was known for sure to be HIV-positive ­ even though the US guidelines for occupational use do not carry such a restriction. PEP would also become more cost-effective if offered only to people whose risk had been substantial, such as unprotected receptive anal or vaginal sex or shared drug injecting equipment - although again, the first paper to propose guidelines for non-occupational use of PEP includes receptive oral sex with ejaculation as sufficient grounds for treatment (Katz & Gerberding). Technology such as viral load testing of the HIV-positive partner might be employed, enabling the prioritisation of cases in which the HIV-positive partner had a high viral load, such as during advanced HIV infection, which might also be expected to increase the risk of transmission. Nevertheless, funding more than the occasional case of PEP for a sexual risk exposure is likely to be beyond the means of most genito-urinary clinic budgets.

Some researchers are concerned that widespread use even of short courses of anti-HIV therapy might encourage the evolution of HIV strains that are resistant to one or more anti-HIV drugs. Resistance may also have implications for the choice of drugs prescribed when PEP is used. Ideally this choice should probably take account of the anti-HIV treatment experience of the HIV-positive partner; for instance, it might be prudent to use an alternative drug such as d4T instead of AZT in cases in which someone has been exposed to an HIV-positive partner who has himself received prolonged AZT therapy and thus may well harbour AZT-resistant virus.

Because of all these concerns, it is probably best if PEP for non-occupational exposures is initially introduced within formal research studies. The advantages of conducting such a study are that it provides a mechanism to collect data on the procedure, and is funded either by a research agency or one or more pharmaceutical company sponsors it may sidesteps the potentially substantial costs. In the absence of any official guidelines on non-occupational use of PEP, a trial protocol also provides a standardised framework for prescribing practice at all the participating clinics.

The first such study is due to begin in San Francisco during 1997. Researchers at the University of California at San Francisco, in collaboration with San Francisco Department of Public Health officials, will publish advertisements encouraging people who believe they were exposed to HIV due to a recent isolated incident of unsafe sex or a condom breakage to come forward to medical services. Those who respond within 72 hours of the exposure will be offered 30 days of treatment with AZT and 3TC; the protease inhibitor indinavir may be added if the source individual is know to have advanced AIDS, a high viral load or prior treatment with nucleoside analogues. Other people who come forward because of previous risks will be offered HIV antibody testing and treatment with anti-HIV drugs if they are found to be HIV-positive.

This study will be too small to assess whether the treatment works, but is expected to produce valuable information about how often these kind of isolated accidental risks (as opposed to regular, planned unprotected sex in relationships, for example) occur, how quickly individuals will come forward afterwards, and what proportion will choose to take PEP once it is offered. The study will also monitor participants' health and sexual behaviour in the long term. It is considered unlikely that a short course of anti-retroviral therapy will cause anything more than short-term side-effects while treatment is under way, although these are sufficient to lead about one-third of health-care workers prescribed AZT monotherapy as PEP to discontinue treatment prematurely; however, unforeseen long-term effects will only be detected by systematic follow-up.


INFECTIVITY

The development of potent anti-HIV regimens that can have a dramatic effect on HIV viral load has coincided with increased interest in whether the infectivity of HIV-positive people varies over time. If there is a relationship between viral load and infectivity, the benefits of drugs that lower viral load might be two-fold ­ benefit to the individual in terms of improved quality and quantity of life, and benefit to society in terms of a reduced risk of transmission to others. Thus, use of anti-HIV drugs might constitute both primary and secondary HIV prevention.

Several research teams have proposed that periods of high viral load in the blood may also be periods of high infectivity. On the basis of complex mathematical models, Jerome Koopman, Professor of Epidemiology at the University of Michigan, has concluded that the vast majority of cases of HIV transmission may occur during the ëprimary infection' stage when the transmitter is himself only recently infected and has very high viral load (Koopman). By analysing data from cohorts of gay men, this team has estimated that the probability of HIV transmission during a single act of anal intercourse is 0.1 to 0.3 (from one in ten to about one in three) during primary infection. At other stages of infection they estimated the risk to be considerable lower: 0.0001 to 0.001 (between one in a thousand and one in ten thousand) during the long asymptomatic period, and 0.001 to 0.01 (between one in a hundred and one in a thousand) once symptoms have developed (Jacque z).

If correct, Koopman's theory has wide-reaching implications. Behaviours that carry at most a modest risk of HIV transmission when undertaken with a partner with established infection, such as oral sex, might plausibly be efficient modes of infection when undertaken with a partner during primary infection. The theory would also lend weight to the importance of early diagnosis of seroconverters. By the time most infected people test positive with an HIV antibody test and thus receive advice or reinforcement of the importance of safer behaviours, the period during which they were most likely to infect others may already have passed. The challenge for HIV prevention campaigns would be to reduce the likelihood of a person who has unprotected sex on one occasion (the source of infection) from having unprotected sex again within the following days or weeks (the high risk period for transmission).

However, the theory assumes that the extremely high levels of viral load seen in the blood of recently infected people are accompanied by a similar peak of viral load in other fluids such as semen. There seems to have been virtually no published research measuring viral load in fluids other than blood during primary infection. However, one recent study on three men in primary infection found that their seminal viral load was higher than the average -- but still within the overall range -- seen among a group of men with established HIV infection (Dyer). The only epidemiological study specifically to examine the impact of primary infection in the real world found no indication of an increased rate of transmission in serodiscordant heterosexual couples during that period (Dueer) . No similar epidemiological studies to assess the significance (or otherwise) of primary infection among gay men have yet been reported.

Nevertheless, even if primary infection is not the sole period of concern, other research has supported a link between viral load and transmission. Some researchers have found a correlation between the level of HIV in the blood and the level of HIV in semen (Coombs ), although others disagree (Jurria ans). A study of people infected with HIV through blood transfusions found that those with higher plasma viral load were more likely to transmit HIV to their partners through heterosexual sex (Lee), and other studies which did not specifically measure viral load have found that unprotected sex with a person with advanced HIV infection, when viral load tends to be higher, carries a greater risk of infection than unprotected sex with an asymptomatic partner (Mastro). Finally, among HIV-positive pregnant women, those with higher viral load are more likely to pass on the virus to their unborn child (Sperli ng).

Thus it remains plausible that reducing viral load with anti-HIV therapy may in turn reduce infectivity. For sexual transmission, this assumes that drugs that reduce viral load in the blood will have a similar effect on virus load in semen and vaginal fluids ­ an assumption supported by some, but not all, of the limited number of studies looking at the effect of anti-HIV drugs on the amount of HIV in semen (Royce ). An Italian epidemiological study found reduced rates of heterosexual HIV transmission from men receiving antiretroviral therapy compared with untreated men (Musicc o). Also, in a placebo-controlled study among HIV-positive pregnant women, the rate of transmission of HIV to their unborn children was indeed reduced from 25% to 8.3% by treatment with AZT during pregnancy, delivery and the infants' first weeks of life. The mechanism of action is thought to be a combination of a reduction in the woman's viral load, and pre- and post-exposure prophylaxis in the child (Sperli ng).

Anti-HIV therapy is by no means the only factor that can influence viral load and thus, probably, infectivity. For example, immunisation with tetanus toxoid temporarily increases viral load in the blood between 2- and 36-fold for up to six weeks (Stanle y). Viral load is also temporarily increased during periods of acute opportunistic infections, as well as during periods of non-HIV-related infections such as sexually transmitted diseases (STDs) (Royce ). In each case, the underlying cause is likely to be immune activation. It is unclear whether such increases would be seen among people taking potent anti-HIV treatments such as combinations that include a protease inhibitor.

It is unlikely that these transient increases in viral load have any measurable impact on the health of the infected person. Their greater significance may be in increasing the infectivity of the individual during risk activities with uninfected people. To the extent that health promotion advice targeted to people with HIV has addressed the issues of STDs, it has usually only warned of their possible effects on the health of the infected individual. While this is perfectly valid, it also seems important to stress that there may be an increased risk of transmitting HIV to others. Sexually transmitted diseases among HIV-negative people have also been associated with increased susceptibility to HIV infection if exposed. Public health interventions that aim to reduce the incidence of STDs both among HIV-positive and HIV-negative people are thus likely to work synergistically in reducing the incidence of HIV transmission.

In 1991 Professor Roy Anderson of Imperial College London and colleagues provoked a brief controversy by arguing that treatment which lengthens the lives of people with HIV could be against the interests of the community as a whole, since by prolonging the period during which HIV-positive people were infectious it could lead to an increased rate of spread of the virus and ultimately a higher total of AIDS deaths (Anders on). However, this model pessimistically assumed that treatments which prolonged life would not also reduce infectivity. If in fact wider uptake of today's potent anti-HIV regimens would lead to a decrease in the rate of HIV spread, it could be considered that both the individual and the public health would be best served by encouraging HIV testing and early treatment.


MOTIVATION FOR SAFER BEHAVIOUR

As the new mood of scientific optimism filters into the public domain, health educators will need to consider carefully the possible effects on individuals' motivation to sustain safer behaviours. In the USA, where many more gay men have now had direct experience of remarkable benefits from the new therapies in their social circles, prevention workers are already expressing concern that the positive developments in treatment may give some men a reason - or simply a new excuse - to abandon safer sex.

To take these fears to their extreme, a well-informed gay man could formulate the following multi-level rationalisation for unprotected sex with a partner whose HIV status he did not know, in which each step is factually correct:

(a) it is statistically improbable that my new partner is HIV-positive, because even in the most sexually active sections of the British gay community, only a minority of men are infected
(b) if he is infected, the average risk of transmission in a single act of unprotected sex is statistically tiny
(c) if he is taking anti-HIV drugs, his infectiousness may well be decreased still further
(d) even if he does pass on the virus, it is possible that I could obtain post-exposure prophylaxis, preventing the infection from becoming established in my body
(e) even if the PEP fails and I do become infected, such prompt use of treatment during seroconversion might dramatically improve my long-term prognosis
(f) even if treatment at seroconversion does not affect long-term outcome, using different combinations of the powerful new anti-HIV drugs may mean that I may still be able to look forward to a very long or perhaps indefinite delay in disease progression.

It would be wrong to assume that this is a major problem facing prevention workers today. It is unlikely that more than a tiny proportion of gay men will subscribe to such an over-optimistic position ­ or indeed that such views would directly lead to unprotected sex anyway. It is more probable that much unprotected sex happens because, frankly, most gay men would rather not use condoms. Afterwards, most of us can probably come up with a long list of 'excuses' that make us feel better about having taken more risks than we think we should, including the effects of drink and drugs. Treatment advances may simply provide a few more 'morning after' excuses.

Nevertheless, it is worth thinking through some of the flaws in the rationalisation. Firstly, many of the more optimistic predictions of the possible effects of treatment remain unsubstantiated. For example, the persistance of HIV DNA in samples from people with undetectable viral load provides solid grounds for suspecting that they remain potentially infectious. Secondly, the effectiveness of post-exposure prophylaxis at preventing sexual transmission remains unproven, and the experience of AZT use by health-care workers shows that the level of protection provided is far from complete, since a proportion do become infected despite receiving PEP. And thirdly, the long-term effectiveness of the new treatments at delaying disease progression and death remains unknown. It is likely to depend on the ability of the drugs to keep HIV replication suppressed indefinitely, plus strict adherence to the guidelines on taking the right dose, at the right time, with or without food as instructed. The disruptive effect that the new treatments can have on daily life ­ not to mention their possible side- effects ­ should leave no doubt that a life-time of HIV therapy is no easy option, compared with simply using a condom. A growing number of people may experience new forms of psychological distress as they anxiously follow trends in their CD4 count and viral load to try to assess whether drug therapy is working. These negative factors are likely to impact most forcefully on asymptomatic people, who might otherwise enjoy a relatively normal lifestyle. It will be important to communicate these harsh realities of the new virology, not just the potential rewards.

Other implications of the medical advances should actively reinforce the importance of avoiding infection with HIV. For example, it is clear that drug-resistant strains of HIV can be transmitted from person-to-person, and it is increasingly common for newly-exposed people to be infected with AZT-resistant HIV (Mayers). It is reasonable to assume that people infected with AZT-resistant HIV will be less able to benefit from treatment regimens containing AZT. They could also have an inherently worse prognosis; several studies have found that the emergence of AZT-resistant strains among people taking the drug is associated with an increased risk of disease progression and death even if they change treatment to ddI (D'Aqui la, Japour ). Infection with strains of HIV that are resistant to other drugs, such as 3TC or nevirapine, has also been seen, and most researchers think that it is simply a matter of time before protease-resistant strains are on the loose.

If medical progress leads to an increase in the number of people with HIV who are on treatment, the proportion of new infections that involves single- or multi-drug resistant HIV strains may rise. Thus, individuals who abandon safer sex because of optimism over the availability of treatments might in practice find themselves unable to benefit fully from those very treatments. (A major unanswered question is whether people who are already infected with a 'wild-type' (drug-susceptible) strain of HIV can be re-infected with a drug-resistant strain, and if so, whether it is likely to replace the wild-type strain and limit their treatment options. In the absence of clear evidence, this possibility is probably the most persuasive reason for suggesting that HIV-positive people should avoid unprotected sex with other positive people.) Increasingly, HIV prevention campaigns will need to take account of these factors in exploring the reasons why risky behaviours persist, and in devising interventions to encourage and support safer sex.


CONCLUSION

It is arguable that there have been few significant developments in HIV prevention approaches since the mid 1980s. The standardised nature of safer sex messages is reflected in the existence of almost formulaic elements in prevention materials ­ familiar statements fossilised for a decade, such as ëIn the absence of a vaccine or cure, safer sex is our only protection against HIV'. Recent developments must force a re-examination of these assertions and the orthodoxy that lies behind them, which too often denies the complexity of the contexts in which decisions about safer sex and other forms of HIV risk reduction are made. The potential contribution of the psychological and social sciences to HIV prevention has long been acknowledged; now, more than ever, it is essential also to recognise the role of medical sciences in illuminating prevention strategies and debates.

To date, HIV prevention has been understood ­ or at least implemented ­ exclusively in terms of individual or community-wide behaviour change. The potential role of a biomedical ësafety net' has been recognised in principle in the recent prioritisation of research into microbicides, which could not only provide a woman-controlled means of HIV prevention when used vaginally, but also expand the prevention possibilities for gay men if used anally (Gorna) . The use of anti-HIV drugs to manipulate viral load and infectiousness may offer a further biomedical approach to HIV prevention, as an alternative or adjunct to conventional behaviour change strategies. Likewise, the old separation between HIV antibody testing and HIV prevention strategies, already seriously challenged by ënegotiated safety' strategies, can only be undermined further as individuals decide to test as a means of access to post-exposure prophylaxis and/or early treatment.

In the USA, treatment education has long been accepted as a crucial service for people with HIV. Many AIDS service organisations have departments devoted to secondary prevention, alongside other care and support provision and primary prevention work.(Watney , 1996). These organisations should be well-placed to assimilate the potential significance of medical advances for primary HIV prevention.

In the UK and much of Europe, however, treatment information has been relatively neglected. In the vast majority of instances, organisations have devoted little or no staff time to any aspect of secondary HIV prevention work, leaving them in the dark about the likely impact of progress in AIDS research on all aspects of their work.(King).

Take another look at that statement: "In the absence of a vaccine or cure, safer sex is our only protection against HIV". What does it therefore mean for safer sex if the prospect of a cure has moved into the realms of the possible?


REFERENCES

Issues in post-exposure prophylaxis and treatment of primary infection. 11th International Conference on AIDS, Vancouver, roundtable session B3, July 1996.

Anderson RM et al. Potential of community-wide chemotherapy or immunotherapy to control the spread of HIV-1. Nature 350:356-359, 1991.

CDC. Case-control study of HIV seroconversion in health-care workers after percutaneous exposure to HIV-infected blood ­ France, United Kingdom, and United States, January 1988 - August 1994. Morbidity & Mortality Weekly Report 44:929-933, 22nd December 1995.

CDC. Update: Provisional Public Health Service recommendations for chemoprophylaxis after occupational exposure to HIV. Morbidity & Mortality Weekly Report 45:468-472, 7th June 1996.

Coombs RW et al. Association between cultivable virus in semen and HIV-1 RNA level in seminal plasma. 11th International Conference on AIDS, Vancouver, abstract We.B.3383, July 1996.

D'Aquila RT et al. Zidovudine resistance and HIV-1 disease progression during antiretroviral therapy. Annals of Internal Medicine 122:401-408, 1995.

Dueer A et al. Risk of HIV transmission during the seroconversion versus the post-seroconversion period. 11th International Conference on AIDS, Vancouver, abstract Mo.C.571, July 1996.

Dyer JR et al. Shedding of HIV-1 in semen during primary infection (letter). AIDS 11(4):543-545, 1997.

Gerberding JL. Management of occupational exposure to blood-borne viruses. New England Journal of Medicine 332:444-451, 1995.

Gorna R. Vamps, Virgins and Victim: How Can Women Fight AIDS? (Cassell, London) pp.303-309, 1996.

Gostin LO et al. HIV testing, counseling, and prophylaxis after sexual assault. Journal of the American Medical Association 271(18):1436-1444, 1994.

Ho DD et al. Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infection. Nature 373:123-126, 1995.

Jacquez JA et al. Role of the primary infection in epidemics of HIV infection in gay cohorts. Journal of Acquired Immune Deficiency Syndromes 7:1169-1184, 1994.

Japour AJ et al. Prevalence and clinical significance of zidovudine resistance mutations in human immunodeficiency virus isolated from long-term zidovudine therapy. Journal of Infectious Diseases 171:1172-1179, 1995.

Jurriaans S et al. HIV-1 viral load in semen versus blood plasma. 11th International Conference on AIDS, Vancouver, abstract Th.A.4036, July 1996.

Katz MH & Gerberding JL. Postexposure treatment of people exposed to the human immunodeficiency virus through sexual contact or injection-drug use. New England Journal of Medicine 336(15):1097-1100, 1997.

King E. Bridging the gap between science and AIDS service provision. In Sherr L, Catalan J & Hedge B (eds.) The Impacts of AIDS: Psychological and Social Aspects of HIV Infection (Harwood, London), 1997a.

King E. Beyond the blood. AIDS Treatment Update issue 50, 1997b.

Koopman JS et al. Core groups cause primary infection to dominate HIV transmission even when more than 90% of virus is excreted during later stages of infection. 11th International Conference on AIDS, Vancouver, abstract Mo.C.570, July 1996.

LaFon SW et al. A double-blind, placebo-controlled study of the safety and efficacy of Retrovir (zidovudine, ZDV) as a chemoprophylactic agent in health care workers. 30th Interscience Conference on Antimicrobial Agents and Chemotherapy, Atlanta, abstract 489, October 1990.

Lee T-H et al. Correlation of HIV-1 RNA levels in plasma and heterosexual transmission of HIV-1 from infected transfusion recipients. Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology 12:427-428, 1996.

Mastro TD et al. Probabilities of sexual HIV-1 transmission. AIDS 10 (suppl A):S75-S82, 1996.

Mayers DL et al. Prevalence and clinical impact of seroconversion with AZT-resistant HIV-1 between 1988 and 1994. 2nd National Conference on Human Retroviruses and Related Infections, Washington, abstract 385, January 1995.

Mellors JW et al. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science 272:1167-1170, 1996.

Moyle G. Resistance to antiretroviral compounds: implications for the clinical management of HIV infection. Immunology and Infectious Disease 5:170-182, 1995.

Moyle G, Gazzard B. Current knowledge and future prospects for the use of HIV protease inhibitors. Drugs 51(5):701-712, 1996.

Musicco M et al. Antiretroviral treatment of men infected with human immunodeficiency virus type 1 reduces the incidence of heterosexual transmission. Archives of Internal Medicine 154:1971-1976, 1994.

Patton C. Media, testing and safe sex education. In Inventing AIDS (Routledge, New York), pp.25-49, 1990.

Royce RA et al. Sexual transmission of HIV. New England Journal of Medicine 336(15):1072-1078, 1997.

Saget BM et al. Dramatic suppression of HIV-1 plasma RNA using a combination of zidovudine, didanosine, zalcitabine, epivir, and interferon-alpha in subjects with recent HIV-1 infection. 11th International Conference on AIDS, Vancouver, abstract We.B.533, July 1996.

San Francisco General Hospital Epi-Center. Post-exposure prophylaxis for occupational HIV exposures. July 1996.

Sperling RS et al. Maternal plasma HIV-1 RNA and the success of zidovudine (ZDV) in the prevention of mother-child transmission. 3rd Conference on Retroviruses and Opportunistic Infections, Washington, late-breaker abstract, January 1996.

Stanley SK et al. Effect of immunization with a common recall antigen on viral expression in patients infected with human immunodeficiency virus type 1. New England Journal of Medicine 334:1222-1230, 1996.

Tokars JI et al. Surveillance of HIV infection and zidovudine use among health care workers after occupational exposure to HIV-infected blood. Annals of Internal Medicine 118:913-919, 1993.

Watney S. Muddling through; the UK responses to AIDS. In Practices of Freedom (Rivers Oram, London) p.249, 1994.

Watney S. The politics of AIDS treatment information activism. In Bywaters P & McLeod E (eds.) Working for Equality in Health, (Routledge, London) 1996.

Wei X et al. Viral dynamics in human immunodeficiency virus type 1 infection. Nature 373:117-122, 1995.

Edward King is an HIV treatment and prevention educator in London,working for the Uk's National AIDS Manual (http://www.nam.org.uk). He also maintains The Mining Company's AIDS/HIV web site at http://aids.miningco.com.