Highly Active Antiretroviral Therapy and Sexual Risk Behavior

---

By Eric Ashton

I. Abstract

II. Introduction (Table 1)

III. Mathematical models of the effectiveness of HAART in the population.

A. The Blower model (Table 2)

B. The factor that HAART decreases transmissibility (Table 3)

C. Relating the Anderson models to the blower model (Equations 1-3)

IV. Is risk behavior increasing due to HAART?

A. Studies that look at reported risk behavior (Table 4)

B. Studies that examine risk-related attitudes about HAART (Table 5)

V. Discussion

I. Abstract

[back to top]

Highly active anti-retroviral therapy (HAART}has the potential to have a positive or negative effect on the HIV epidemic in a community. While the drugs may decrease the transmissibility of the virus, they allow people with HIV to live longer and have the chance to infect more people. Also, there is evidence that the presence of HAART may lead to increased sexual risk behavior. Mathematical models have been created to attempt to quantify these effects and see how they relate to one another. These mathematical models are examined and the studies that indicate sexual risk behavior may be on the rise are reviewed thoroughly. The implications of these studies are discussed and recommendations are made.

II. Introduction

[back to top]

Mortality and morbidity from HIV disease in the United States has declined sharply in the last 4 years. This reduction was caused by the widespread use of highly active antiretroviral therapy (HAART), which started in about 1996 [18]. The term HAART refers to combining nucleoside analog drugs such as zidovudine (AZT) with non-nucleoside analog reverse transcriptase inhibitors and protease inhibitors such as saquinavir, ritonavir, indinavir, or nelfinavir. HAART has been shown to be a safe, effective, and relatively cost-effective way to decrease AIDS incidence and death rate from HIV [7, 18] It has been hypothesized that the mediating factor in the efficacy of HAART is the lowering of the level of virus in the blood. HAART often lowers the level of viral RNA present in the blood below the minimum level that is detectable by the most sensitive assays (<20 copies/mL) [7].

Despite the overwhelming evidence that HAART is helpful on an individual basis it is possible that it may have negative effects on society as a whole [1, 10, 2]. For example, the presence of HAART may lead to a higher level of sexual risk behavior that might detract from or even offset the benefits of the drugs. Table 1 summarizes reasons why HAART may have a positive or negative effect on transmission of HIV:

Possible Positive Effects of HAART	Possible Negative Effects of HAART
HAART may decrease the per-person transmission risk. [17]	Since people with HIV will live longer they have the opportunity to transmit the virus to more people [10]
The presence of effective therapy for HIV disease may lead to more people being tested and thereby have lower risk behavior. [9]	Increased sexual risk behavior per person per year due to [16, 6] Better health An attitude that AIDS is 3under control4 Belief that they are less likely to transmit Less reminders of disease since morbidity and mortality is going down
Drug-resistant strains of HIV might be harder to transmit since they have demonstrated a lower replication capacity in vitro [15]	Resistance will lead people to stop taking the drugs which then leads to primarily non-resistant virus circulating in blood [6]. This might mean that the per-person transmission risk lowering is only temporary.

Table 1 6 Reasons why HAART might increase or decrease HIV transmission

III. Mathematical models of the effectiveness of HAART in the population

[back to top]

A. The Blower model

A recently published study by Blower et. al. [2] looks directly at the issue of the affect of HAART on the HIV epidemic at the societal level. They made a mathematical model of the HIV epidemic among gay men in San Francisco. This was a Monte Carlo model; they ran it 1000 times with different sets of parameters randomly selected within ranges that the authors believed best reflected the real epidemic in San Francisco. The purpose of this model is to see what would happen if the usage rate of HAART went up from the current level 50%, to levels between 50% and 90%. They made separate optimistic and pessimistic analyses. In the optimistic analysis, the only experimental variable was the usage rate of HAART. That is, they compared the HIV incidence and death rate in models assuming that HAART usage was between 50-90% and what the same rates would be if there was no HAART therapy. In the pessimistic analysis there was an additional experimental variable 6 the increase in sexual risk behavior. They compared the death and incidence between a world with a high usage of HAART and a certain increased level of sexual risk behavior and a world with no HAART and a consistent level of 1.7 risky sex partners per year. Table 2 lists the parameters of the model that relate most directly to the effect of HAART on HIV incidence:

Variable	Range of values selected:
Number of risky sex partners per year (optimistic)	1.7
Factor of increase in sexual risk behavior (pessimistic)	No-effect to 2-fold (median 1.5 fold)
Number of risky sex partners per year (pessimistic)	1.7 - 3.4
Years from infection to death of an untreated man with HIV	12
Factor of increase of the lifespan of a man treated with HAART	1.5-fold to 3-fold (median 2.2 fold)
Years from infection to death of a man treated with HAART	18 - 36
Chance of contracting HIV per untreated partner	10%
Factor transmissibility is reduced by HAART treatment	2-fold to 100-fold (median 3.9 fold)
Chance of contracting HIV per HAART treated partner	0.1% - 5%

Table 2 6 Assumptions of the Blower model [2]. Values were sampled 1000 times from a uniform probability function (PDF). The median for the third factor is skewed because the values were sampled uniformly from the reciprocal of that factor (0.01 to 0.5 median = 0.255)

In the optimistic analysis the presence of HAART lead to decreases in the death rate from HIV and the incidence. In the pessimistic analysis the presence of HAART and the increased sexual risk behavior still led to a short-term decrease in the death rate however they there was an increase in the incidence of HIV. The authors conclude that 3Increasing the usage of ART in San Francisco would decrease the AIDS death rate and could substantially reduce the incidence rate.4 and that the usage of HAART 3should4 be increased in San Francisco. [2]

However there are some inconsistencies which calls their conclusions into doubt. Using the median assumptions of their model a person who develops resistance to HAART usually gives up treatment at the rate of 26% per year. After six weeks of being off treatment they revert back to being drug sensitive and can be treated effectively again. [2] This result is inconsistent with clinical evidence that people who develop resistance to antiretroviral drugs, stop using them for some time, then go back on them get no therapeutic effect from them. It is true that the resistant virus is not typically the dominant strain after time passes but if the therapy is started again the resistant strain is quickly selected for again. [6]

Also there have been studies that have shown that an alarmingly high proportion of new HIV infections are resistant to antiretroviral drugs. One study found this proportion to be 16% in New York [5] another found it to be 11% in Geneva, Switzerland [21]. Using the median assumptions of the Blower model only 10% of patients develop resistance every year and that resistance only lasts for about 4 years. During that time they are only half as likely to pass on the virus due to the decreased fitness of the virus. Since they are still taking the drugs for most of the time their risk of transmission is cut in half again. Even if they do manage to pass on the virus there is only a 50% chance they will transmit drug-resistant virus [2]. If these assumptions were close to being correct, then there would be so many factors working against new infections being resistant that a quick increase in the percent of resistance in primary infection would not be possible. However, a quick increase is exactly what seems to be happening [5, 21].

B. The factor that HAART decreases transmissibility

[back to top]

Another weakness of the Blower logic is their explanation for why HAART is expected to be so effective at reducing the per-partner transmission rate. The authors argue that AZT has been shown to decrease transmission risk by decreasing viral load and that since HAART is 10- to 100-fold more effective than AZT at decreasing viral load they expect HAART to cause a 2- to 100-fold decrease in transmission risk. This makes the assumption that viral load is inversely proportional to transmission risk [2].

The study that most convincingly shows the link between decreased viral load and transmission risk is the Rakai study which looked at 415 HIV discordant couples in rural Uganda where antiretroviral treatment is not available [19]. This paper was not referenced by the Blower article because it was published two months afterwards. The Rakai researchers did viral load tests from the HIV-positive member of the discordant couples. They found that couples where the HIV-positive member had lower viral loads were less likely to transmit. The primary conclusion of their study was that 3The viral load is the chief predictor of the risk of heterosexual transmission of HIV-1 and transmission is rare among persons with levels of less than 1500 copies per mL4 [19]. .

However, it is possible that there are confounding factors other than viral load that correlated with both high viral load and higher transmission risk. Perhaps the immune systems of some of the people were less effective at suppressing the virus because they were also dealing with sexually transmitted diseases which might have made transmission more likely. Also, viral load is naturally high at the beginning of HIV infection before the immune system starts to effectively fight the virus 6 this beginning time is also likely to correlate with high risk behavior since that they are probably continuing the high level of risk that led them to get the virus. Also there might have been genetic or other immunologic factors that may have caused viral load and transmission risk to go down that would not be repeatable by giving HAART. One of the biggest points of the study was that among the 51 patients with naturally-occurring abnormally low viral loads (below 1500 copies) there was little transmission. However, that does not necessarily mean that artificially lowering someone2s viral load to that level will have a similar effect on transmission.

A study that had similar methodology and conclusions was done in the United States and is the one referenced by the Blower article as the evidence that AZT has been shown to lower the risk of sexual transmission by reducing viral load. In this study [17] there were 436 discordant couples of which 15% were on AZT therapy (all of these couples were male-positive). Because only half as many of the men treated with AZT transmitted the virus as untreated men the authors concluded that 3Treatment of HIV infected men with AZT reduces, but does not eliminate, heterosexual transmission of infection.4 However, this causal relationship is not proven by their data. There may have been many factors that confounded the results. The 15% on AZT therapy were not randomly selected and it seems likely that the group may have been richer, more socially conscious, or more trusting of their doctors which might have led them to have less risk behavior 6 a factor which was not asked or accounted for [17].

Even if HAART was completely ineffective in stopping transmission, we might still expect to see a repeatable correlation between viral load and transmission. Table 3 is a conceptual description of events in the course of infection that may be associated with changes in both viral load and sexual risk behavior:

Event	Expected Viral Load	Expected Risk Behavior
Person first infected	High	High
The persons immune system controls the virus level and the person gets bored with high level of sex risk that, unbeknownst to him, got him infected.	Medium	Medium
Person finds out they have HIV and starts HAART	Very Low	Very Low
HAART becomes less effective through time, as do messages about safer sex	Raises with time	Raises with time

Table 3: A conceptual framework to explain why transmission risk might correlate with viral load even if HAART did not itself lead to lower transmission.

The idea that HAART decreases per-partner transmission is supported by a number of studies that have shown less virus in semen of men on HAART. [reviewed in 11] However, all of these studies just showed short term results and it will be necessary to do longer-term follow-ups in order to asses the duration of the effect [11]. There is also evidence that there are limits to the amount that lowering blood viral load will lower the amount of virus in the semen. Zhang et. al. [22] has shown that even in men with undetectable viral loads there is virus in the semen which indicates transmission is still possible.

C. Relating the Anderson models to the Blower model

[back to top]

In order to examine conceptually how the key variables of the Blower model (Table 2) relate to the affect HAART has on HIV incidence we will look to the work of Roy Anderson who worked on mathematical models published in 1991 and 1996 which quantify the possible effects of antiretrovirals on the HIV epidemic [1,10]. The 1991 paper [1] presents equations that explain how AZT (the only antiretroviral available at the time) might have a negative effect on controlling the spread of HIV. The authors conclude: 3Community treatment with antiviral drugs or immunotherapies that lengthen the incubation period of AIDS without significantly reducing the infectiousness of treated individuals, can increase the rate at which HIV-1 infection spreads (which is fairly obvious) and can even, under certain circumstances, increase the AIDS-related death rate in the community (which is less obvious).4 [1] The 1996 article [10] details the basic equations much more and shows the relationship between the different factors that can affect the effect that antiviral therapy has on the community. Equation 1 is the basic equation used in that article This equation explains the spread of a sexually transmitted disease in a behaviorally homogenous population (like the one Blower assumes):

(Equation 1, from[10])

R₀ is the basic reproductive number which is the average number of secondary cases of infection generated by one primary case. This is determined by multiplying the duration of the infectious period, D, by the number of sex partners per year, c, and the per-partner chance of transmission, b. If R₀ < 1 then the epidemic will die out, if R₀ > 1 then it will spread. If introducing a new therapy causes R₀ to go down then it will slow the spread of disease if the new therapy causes R₀ to go up then the new therapy will cause the epidemic to grow. [10]

In order to relate these variables to the ones from the Blower model (Table 2) it will be necessary ask what will introducing a new therapy to everyone in the community will do them. If the entire community were to start using HAART D, c, and b might all change as a result. Therefore it will be useful to distinguish between these parameters in the untreated world D_U, c_U, and b_U and in the treated world D_T, c_T, and b_T. The treatment is slowing the epidemic if the treated value of R₀ is less than the untreated value, that is if the following relation is true:

(Equation 2)

Rearranging these variables to get positive ratios gives:

(Equation 3)

If HAART were introduced to everyone with HIV in a community then these three terms would be the same as the three 3factors4 from the Blower model (Table 2). D_T/D_U is the factor of increase of the lifespan of a man treated with HAART, c_T/c_U is the factor of increase in sexual risk behavior, and B_U/B_T is the factor transmissibility is reduced by HAART treatment. To see if implementation of the therapy is helping transmission just multiply the two 3bad4 factors (c_T/c_U, D_T/D_U) together and hope that it is lower than the one 3good4 factor (B_U/B_T). In order to adjust this formula for less than perfect implementation of the therapies simply modify D_T, c_T, and b_T to reflect which percentage of HIV-positive people are expected to be on the therapy.

There are no studies currently available that quantitatively compare transmissibility with and without (or even before and after) HAART. Also the factor that HAART increases average lifespan is impossible to quantify at this point since HAART has only been used for about 5 years. The remainder of this review focuses on the possibility of increase of sexual risk behavior due to the introduction of HAART.

IV. Is risk behavior increasing due to HAART?

[back to top]

There are a number of reasons why HAART might theoretically lead to an increase in sexual risk behavior. People on HAART may be healthier, which might lead to an increased sex drive [16]. The introduction of HAART may have caused an attitude shift in society as whole to believe that AIDS is not as big a worry now as before [13]. People on HAART may feel that they are less likely to transmit the virus [14]. Since morbidly and mortality are declining then the reminders people see of HIV disease will be less [18]. Those reminders may play an important role in risk reduction

A number of surveys have been published recently that indicate that the rate of sexual risk behavior has been increasing in places where HAART use became widespread in 1996. Table 4 summarizes the methodology of these surveys.

Reference	Location	Sample	Timing	Data
CDC [4] 1999	San Francisco	21857 MSM recruited in non-medical settings.	6223, 5989, 5472, and 4173 men in each year from 1994-1997.	One-page written questionnaire that assessed use of condoms and types of sex behavior
CDC [4] 1999	San Francisco	City-wide rate of rectal gonorrhea in MSM	Measured yearly from 1990-1997	Shows trends in rate of RG throughout the city
Dodds [9] 2000	London	6671 homosexual men recruited from commercial gay venues and STD clinics	2482, 2121, 2068 men in each year from 1996-1998	Anonymous questionnaire that assessed unprotected sex and sex with HIV discordant persons
Miller [16] 2000	French SEROCO group	191 HIV-positive individuals who initiated HAART after Jan 1996	Compares responses to interview before and after starting HAART	Interview that assesses the types of sex, the status of the partners and the use of condoms
Dilley [8] 1997	San Francisco	54 sexually active MSM who were going in for HIV testing	Given in 1996-7. Results not time dependant.	Asked about attitudes concerning treatment and risk
Kelly [13] 1998	Milwaukee	379 MSM who were aware of HAART recruited from gay pride festival, gay newspaper ad, and AIDS service organization	Given mid-1997 to early 1998. Results not time dependant	Asked about attitudes concerning treatment and risk
Kravcik [14] 1998	Ottawa	147 HIV-positive individuals at HIV clinic of Ottawa general hospital	Given in 1997 Results not time dependant	Asked about attitudes concerning treatment and risk

Table 4 6 Methodology of surveys that show significant increase in sexual risk behavior or in risky attitudes.

MSM = men who have sex with men (a term that refers to homosexual men and bisexual men as a group).

All of these studies either pointed to a significant increase in reported risk behavior in the time since HAART has become available [4, 9, 16] or showed a significant presence of attitudes that support the idea that HAART might lead to increased sexual risk behavior [8, 13, 14].

A. Studies that look at reported risk behavior

[back to top]

The San Francisco and London surveys were large-scale population surveys of HIV positive and negative MSM in the entire city [4,9]. The CDC questionnaire found that the proportion of men in San Francisco who reported having unprotected anal intercourse increased from 24% in 1994 to 33% in 1997. The proportion of men reporting 3always4 using condoms decreased from 70% to 61%. These results were corroborated by an increase in the rates of rectal gonorrhea which had been going down previous to 1993 but increased each year from 1994-1997, almost doubling in that period with the largest jump in 1996 [4]. The London survey found similar increases over the years 1996-1998. The percent reporting unprotected anal intercourse with more than 1 partner in the last year jumped from 32% to 38%. The percent who reported a sexual partner or unknown or discordant HIV status in the past year jumped from 18% to 21% [9]. .

While both of these studies show increases in sexual risk happening at about the same time as the introduction of HAART there may be other reasons for the increase. For example, as time goes on the attitudes instilled by education and prevention campaigns may wane naturally. However, it would be quite a coincidence if the increase was completely independent of the media coverage about HAART.

The French study [16] differed from the ones above since it was limited to HIV positive patients. It compared survey results from the same people taken at different times 6 the year before they initiated HAART therapy and the year after. Among heterosexual men and women there were slight decreases in the amount of reported risk behavior, however these differences were not statistically significant. In MSM there were increases found in sexual risk behavior. The only difference that was statistically significant was the reported unprotected sexual intercourse that MSM had with partners who were of HIV-negative or unknown status. 3 out of 65 reported such encounters in the year before starting HAART, 9 out of 65 reported such encounters in the year afterwards [16].

While the French study does show a significant correlation between HARRT use and one type of sexual risk it is not possible to conclude a causal relationship based on their data. There are many confounds that were not controlled for that might explain the differences in response at the later survey time. At the later time the group would have known their HIV status for longer, the date would be significantly later, and they would have been given the same interview multiple times. Maybe all this study is showing is that if you ask gay men the same questions about sex repeatedly they get more comfortable telling the truth while heterosexual men and women get less comfortable. Also, the interviewers were not blind to the treatment status or sexual orientation of the subjects [16].

B. Studies that examine risk-related attitudes about HAART

[back to top]

In order to look more directly at the question of the affect of HAART on people2s attitudes about safe sex, a number of surveys have specifically asked people questions about HAART and risk that are designed to assess the prevalence of attitudes that might lead to increased risk behavior. All of these surveys indicate a small but significant presence of attitudes that might link HAART to increased risk behavior. [8, 13, 14].

The Dilley et. al. [8] study looks directly at the question of whether advances in treatment are changing views about high risk sex. Of the 54 San Francisco MSM , 26% agreed with the statement 3Because of the new treatments for HIV-positive people I am less concerned about becoming HIV-positive myself4 13% indicated that they were more willing to take a sexual chance now because of the new treatments and 15% indicated that they had already taken such a chance due to the presence of the treatments. 17% of men agreed with the statement that the new treatments make you less likely to transmit, 76% of men disagreed with it and 7% indicated that they did not know [8].

The Milwaukee study [13] looks at the question from a slightly different perspective. Their questions were directed at the perceptions of the severity of the AIDS epidemic and the need to maintain safer sex. 379 MSM were surveyed and 10% agreed with the statement 3AIDS is now very nearly cured.4 16% of respondents admitted that they would stop using condoms altogether if the cure for AIDS was announced. Of the 50 of these men who are HIV-positive and on HAART, 18% indicated that they practiced safer sex less often since the new treatments came along. Unlike the Dilley study [8] they only asked if they their practices had changed 3since4 the introduction of the new therapies, not 3because4 of the therapies [13].

The Kravick et. al. survey [14] looks at the attitudes of HIV-positive individuals in Ottawa. It attempted to see if people thought that taking HAART drugs made certain sex practices more or less risky or if being on treatment made people think safer sex was less important. Each respondent was presented with three hypothetical HIV discordant couples where the man was HIV positive and his partner was HIV negative.

1) 3No therapy4 The man is HIV positive and is taking no antiretroviral agents and has a T-cell count of about 200 cells/mL.

2) 3RTI therapy4 The man is HIV positive and is taking ZDV and Lamivudine and has a T-cell count of about 200 cells/mL.

3) 3PI therapy4 The man is HIV positive and is taking ZDV, Lamivudine, and a protease inhibitor, has a T-cell count of about 200 cells/mL, and whose viral load is 3undetectable4

Next, each respondent was asked to fill out a questionnaire 6 part of this is reconstructed in Table 5. This is not exactly what their survey looked like 6 it is a reconstruction of it based on their description [14].

How risky would for each of these couples to engage in anal sex?
No therapy	Very risky		Moderately risky	Somewhat risky	Little risk		No risk at all
RTI therapy	Very risky		Moderately risky	Somewhat risky	Little risk		No risk at all
PI therapy	Very risky		Moderately risky	Somewhat risky	Little risk		No risk at all
How important is it that each of these couples practice safe sex for anal sex?
No therapy	Very important	Moderately important		Somewhat important	Not very important	Not at all important
RTI therapy	Very important	Moderately important		Somewhat important	Not very important	Not at all important
PI therapy	Very important	Moderately important		Somewhat important	Not very important	Not at all important

Table 5 6 A reconstruction of part of the questionnaire used by Kravick et. al. [14].

These same questions were repeated for vaginal sex, oral sex, needle sharing, and blood donation. They found a small but significant proportion of people thought that PI therapy made sex less risky and safer sex less important. 99% responded that no-therapy anal sex was 3very risky4 but only 91% of respondents put 3very risky4 for anal sex for the PI therapy couple. More than 20% of respondents believed that PI therapy reduced the risk of at least one activity [14].

These three attitude surveys [8, 13, 14] all show an small but significant presence of alarming attitudes about HAART, sex, and the status of the AIDS epidemic. However, it seems likely that they may be underestimating or overestimating the actual presence of such attitudes. Since the people giving these surveys were health care workers and public health personnel there is probably a significant bias in the data towards what the respondent thought the person wanted to hear. If someone is working in and AIDS education and prevention capacity asks a person if he think AIDS is not a worry anymore or a person who is giving out condoms asks if a person if he thinks safer sex is still important then nature of the survey givers work might influence the results. Also, there is a significant possibility that the phrasing of the questions may lead people to come to conclusions that they may really believe. For example, if someone is presented with the questionnaire in Table 5, it may seem like the questions are designed to produce a different response for the three couples. In the Dilley study [8] only 7% of MSM surveyed responded 3Don2t know4 to the statement 3I am less likely to get infected from a guy who is on those new treatments than from a guy who is not on them.4 Since this percentage is so low it indicates that respondents might have been jumping to conclusions 6 either second-guessing the question by thinking that the San Francisco Department of Public Health workers wanted or expected them to respond 3disagree4 (76% disagreed) - or assuming that because they are being asked the question that it might be reasonable to agree with the statement (17% agreed) [8].

Another possible reason that antiretroviral may lead to increased sexual risk behavior is the increasing availability of post-exposure prophylaxis (PEP) for HIV. At first, such treatment was given to health care workers who were accidentally exposed to HIV in the course of their work. Retrospective studies have shown that PEP is associated with lower risk of transmission in health care workers [3] Increasingly services are arising that will make the drug available after risk sex encounters. There is a danger that, in knowing such treatment is available, people will engage in more sexual risk. [12].

Another factor that may lead to increased sexual risk behavior is the common practice of telling patients on HAART that their viral load is 3undetectable4 A survey of 104 HIV discordant couples from around California found that there was significantly more reported sex between couples who were told that the HIV-positive person had an 3undetectable4 viral load [20].

V. Discussion

[back to top]

The weight of the evidence seems to suggest that sexual risk behavior is going up 6 while it can never be proven that this would not happened without the introduction of HAART. It is unclear whether the effect is enough to offset the benefits of HAART in reducing HIV incidence 6 to answer this question we will require a more concrete understanding of the effect HAART has on transmissibility and the average time that HAART increases lifespan.

The Blower model attempts to estimate these factors as well as we know and comes up with mixed results 6 indicating that HAART may increase or decrease the HIV incidence rate depending on which assumptions are used. It found that an increase in sexual risk behavior of only 10% would be enough to outweigh the benefits of HAART in terms of HIV incidence [2]. One surrogate for measuring the rate of unprotected sex among gay men in San Francisco is the rate of rectal gonorrhea, which has almost doubled in the time from 1993-1997 [4].

The 1991 Anderson article points out that even if antiretroviral therapy was found to be bad for the community as a whole it is not ethical to deny a treatment is effective on an individual basis [1]. Therefore, it is more useful to think of these results not in terms of whether HAART is good or bad, but in terms of underlining the need for increased vigilance. Clearly, getting the message out about safer sex is more important now than ever. Also, a review of articles presented in the popular media in 1996 should be conducted to see if there was a irresponsible level of sensationalism.

It is still common practice for doctors to use the term 3undetectable4 to describe the viral load of patients who have newly started HAART [14, 20] . However, using this word is not only irresponsible but inaccurate as well. The reason the virus is not picked up on the tests is because the tests used are not sensitive enough. [7]. The viral proviral DNA and viral RNA is still there and there is no reason to believe it is technically impossible to detect them. In light of evidence specifically linking the use of the word 3undetectable4 to increased sexual risk behavior [20] it seems imperative that doctors resist the temptation to provide false hope and instead use a more technically accurate and likely less risk-inducing term such as 3below threshold.4 In explaining the results of the tests, perhaps doctors should emphasize with their patients the possibility passing on resistant strains. That way, when someone comes back with a lab test negative for HIV RNA they can be instructed that they are putting potential sex partners at the risk of not only getting HIV, but of getting drug resistant HIV. If the patient asks the doctor about whether they are less likely to transmit the virus because of their low viral load the doctor should point to the evidence which shows a correlation, but not necessarily causation, between therapy, low viral load, and less transmission [17, 19].

References

[back to top]

1. Anderson, et. al. 3Potential of community-wide chemotherapy or immunotherapy to control the spread of HIV-1.4 Nature 350:356-359. 1991. [ABSTRACT]
2. Blower, et. al. 3A tale of two futures: HIV and antiretroviral therapy in San Francisco.4 Science 287: 650-654. 2000. [FULL TEXT]
3. Centers for Disease Control and Prevention. 3Case-control study of HIV seroconversion in health-care workers after percutaneous exposure to HIV-infected blood 6 France, United Kingdom, and United States, January 1988-1994.4 MMWR. 44:929-933. 1995. [FULL TEXT]
4. Centers for Disease Control and Prevention. 3Increases in unsafe sex and rectal gonorrhea among men who have sex with men 6 San Francisco, California 1994-1997.4 MMWR 48:45-48. 1999. [FULL TEXT]
5. Daniel et. al. 3HIV-1 drug resistance in newly infected individuals.4 JAMA 282:1135-1141. 1999. [FULL TEXT]
6. Davenport 3Reversibility of HIV drug resistance.4 Science 288: 1299. 2000. [FULL TEXT]
7. Deeks et. al. 3HIV-1 protease inhibitors: a review for clinicians.4 JAMA 227:145-153. 1997. [ABSTRACT]
8. Dilley et. al. 3Are advances in treatment changing views about high-risk sex?4 NEJM 337: 501-502. 1997. [FULL TEXT]
9. Dodds et. al. 3Increase in high risk sexual behaviour among homosexual men, London 1996-8: cross sectional, questionnaire study.4 BMJ 320: 1510-1511. 2000. [FULL TEXT]
10. Garnett and Anderson 3Antiviral therapy and the transmission dynamics of HIV-1.4 Journal of Antimicrobial Chemotherapy 37(Supl. B): 135-150. 1996. [ABSTRACT]
11. Kashuba et. al. 3Antiretroviral-drug concentrations in semen: implications for sexual transmission of HIV-1.4 Antimicrobial Agents and Chemotherapy. 43:1817-1826. 1999. [PROMED REFERENCE]
12. Katz et. al. 3Postexposure treatment of people exposed to HIV through sexual contact or injection-drug use.4 NEJM. 336: 1097-1099. 1997. [FULL TEXT]
13. Kelly et. al. 3Protease inhibitor combination therapies and perceptions of gay men regarding AIDS severity and the nee to maintain safer sex.4 AIDS 12:F91-F95. 1998. [ABSTRACT]
14. Kravick et. al. 3Effect of antiretroviral therapy and viral load on the perceived risk of HIV transmission and the need for safer sex practices.4 Journal of AIDS and Human Retrovirology. 19:124-129. 1998. [ABSTRACT]
15. Mammano et. al. 3Resistance-associated loss of viral fitness in HIV-1: phenotypic analysis of protease and gag coevolution in protease inhibitor-treated patients.4 Journal of Virology. 72:7632-7637. 1998. [FULL TEXT]
16. Miller et. al. 3Sexual behavior changes and protease inhibitor therapy4 AIDS 14:F33-F39. 2000. [ABSTRACT]
17. Musico et. al. 3Antiretroviral treatment of men infected with HIV-1 reduces the incidence of heterosexual transmission.4 Achieves of Internal Medicine 154:1971-1976. 1994. [ABSTRACT]
18. Palella et. al. 3Declining morbidity and mortality among patients with advanced HIV infection.4 NEJM 338: 853-860. 1998. [ABSTRACT]
19. Quinn et. al. 3Viral load and heterosexual transmission of HIV-1.4 NEJM: 342: 921-929. 2000. [ABSTRACT]
20. Vanderstraten et. al. 3Sexual risk behaviors among heterosexual HIV serodiscordant couples in the era of post-exposure prevention and viral suppressive therapy4 AIDS 14: F47-F54. 2000. [ABSTRACT]
21. Yerly et. al. 3Transmission of antiretroviral-drug-resistant HIV-1 variants4 The Lancet. 354: 729-733. 1999. [ABSTRACT]
22. Zhang et. al. 3HIV-1 in the semen of men receiving highly active antiretroviral therapy.4 NEJM 339:1803-1809. 1998. [ABSTRACT]