Predictors of the use of viagra, testosterone, and antidepressants among HIV-seropositive gay and bisexual men : AIDS

Secondary Logo

Journal Logo

Original papers

Predictors of the use of viagra, testosterone, and antidepressants among HIV-seropositive gay and bisexual men

Purcell, David Wa; Wolitski, Richard Ja; Hoff, Colleen Cb; Parsons, Jeffrey Tc; Woods, William Jb; Halkitis, Perry Nd

Author Information
AIDS 19():p S57-S66, April 2005. | DOI: 10.1097/01.aids.0000167352.08127.76
  • Free



Recent data in North America, Australia, and Europe suggest a resurgence of unprotected sex, sexually transmitted diseases (STD), and HIV infections among gay and bisexual men [1–4]. Despite many potential explanations for these trends, substance use is one factor frequently found to be associated with sexual risk behavior and the acquisition or transmission of STD or HIV among gay and bisexual men [5,6]. However, most of the literature has focused on the relationship between illicit substance use and sexual risk [6–9]; fewer studies have focused on the use of prescription drugs that may affect sexual behavior, particularly among HIV-positive men.

Our study focuses on the use of sildenafil citrate (viagra), testosterone, and antidepressants among a large, diverse sample of HIV-positive gay and bisexual men recruited for an HIV prevention intervention trial. These drugs are used commonly to treat medical conditions in HIV-positive men, but they also have direct or indirect effects on sexual behavior. It is important to understand how HIV-positive gay and bisexual men use these drugs, whether they are taken in ways that are contraindicated, who uses them, and whether men who engage in sexual risk practices are more likely to use these drugs.


Viagra was introduced in 1998 as a temporary or episodic pharmacological treatment of sexual or erectile dysfunction (ED) and its popularity quickly soared [10]. Shortly after taking viagra, a man is able to develop and maintain an erection if sexually stimulated, but the effects wear off in hours. The prevalence of ED is estimated to be 10% in the general population, although ED increases with age and with the presence of a variety of medical conditions, including HIV [11,12]. The prevalence of ED among HIV-positive gay men is estimated to be 33%, and ED increases with more advanced HIV disease [13]. Medical factors that may contribute to ED among HIV-positive men include the side-effects of prescribed medications as well as medical conditions such as depression, endocrine abnormalities, neurological disorders, and many others [11,14–16]. ED also can be caused by psychological factors, which for HIV-positive men may include the fear of infecting a partner, shame, and anxiety about disclosure.

Viagra is contraindicated with protease inhibitors and with nitrates. HIV-positive men who take protease inhibitors need to use viagra cautiously or in lower, more widely spaced doses, because some protease inhibitors, particularly ritonavir, slow the metabolism of viagra, leading to higher concentrations and a greater likelihood of side-effects [11,17]. Nitrates are commonly prescribed for heart disease, but in a different form, called poppers (amyl nitrates), they are one of the most commonly used illicit drugs by gay men, and their use is frequently associated with sexual risk [9,18]. The use of viagra and drugs containing nitrates is contraindicated because each drug can decrease blood pressure, and the combination could lead to a stroke or myocardial infarction [17,19].

Viagra use appears to be increasing among gay men; its use is associated with illicit substance use, and it is often used without a prescription. Studies conducted among gay men in 1999 reported lower rates of use (12.0–14.8%) [20,21] than studies conducted a few years later, sometimes in the same locales (31–32%) [22,23]. Among STD clinic patients in San Francisco, the rate of viagra use in the past 12 months was much higher by gay and bisexual men (31%) than by heterosexual men (7%) [23]. A study with a more representative sample of gay and bisexual men in San Francisco [22] found that 32% of the men had ever used viagra, 21% had used viagra in the past 6 months, and recent use was more common among HIV-positive men (42%) than among HIV-negative (19%), or unknown-serostatus (12%) men. Data show that more than one third of men combine viagra with other substances, and some men use it to counteract the effects of substances, such as methamphetamines, that can cause ED [20–24]. Men who mixed viagra with other drugs, compared with men who took viagra alone, were more likely to be younger, to get viagra from a friend, to believe that viagra and other drugs combined would enhance sex, to report more recent sex partners and more sexual risk [23]. Men who use viagra have also been found to be more likely to use anabolic steroids [21]. Of the men taking viagra, consistently less than half (17–44%) reported that it was prescribed for them [21–23], increasing the likelihood of misuse and abuse. One study found that HIV-positive men were just as likely to have taken viagra without a prescription as were HIV-negative men [21].

Viagra use among HIV-positive gay and bisexual men is also important to public health because its use has been associated with unprotected sex [20–23]. Two studies suggest that viagra use is associated with unprotected serodiscordant sex by HIV-negative men, but less so by HIV-positive men, although the sample sizes of HIV-positive men were small [21,23]. However, HIV-positive men who used viagra were almost twice as likely to have an STD (50%) than were HIV-positive men who did not use viagra (26%) [23]. Our study can help determine whether HIV-positive men who use viagra, compared with men who do not, may be engaging in more unprotected sex with HIV-positive partners (thus being vulnerable to STD), but not significantly more unprotected sex with serodiscordant partners, as has been suggested by previous research.

Testosterone and other anabolic steroids

A number of different anabolic steroids are prescribed for HIV-positive individuals to help decrease fatigue and reverse cachexia (or ‘wasting’), with the most common being testosterone [25,26]. Because of their abuse in the bodybuilding and sports world, anabolic steroids were banned for general use in 1990, and were designated class III controlled substances [26]. Treatment with testosterone increases abnormally low testosterone levels and has both anabolic effects (increased lean body mass and muscle strength) and androgenic effects (improved libido and energy). Unlike viagra, which has an immediate, short-term effect, steroids are related to short and long-term effects such as an improvement in sexual interest (libido), erectile functioning, and satisfaction with one's sex life among HIV-positive men [27].

Given the focus on physical strength and sexual prowess in the gay community and among HIV-positive men [28], and the shift in cultural norms regarding the ideal male body towards increased muscularity [29], some men are strongly tempted to abuse steroids. For example, more than 13% of gay men recruited in London gyms in 1997 reported lifetime use of steroids, but only 36% of these men had discussed their use with a physician [30,31]. In 2000, 15% of men in London gyms reported the use of steroids in the past year, with HIV-positive men more likely to report use (32%) compared with HIV-negative men (15%) or untested men (5%) [32]. Among a small sample of HIV-positive men recruited over the internet (N = 114), 33% of participants reported using testosterone, 17% used another steroid (deca durabolin), and 16% used both, with 36% of participants reporting that they use steroids to ‘pump up’ [33].

Regarding HIV risk behavior, needle sharing appears to be very rare among anabolic steroids users, and testosterone is also available in other forms, including a patch or gel [17,32,34]. However, the use of anabolic steroids is related to unprotected sex among gay men [30], among male high school students [35], and among weight trainers [36], although some of these associations may be a result of overall risk-taking personalities rather than something specific about anabolic steroids and sexual risk. Overall, anabolic steroids provide benefits to many HIV-positive men, but they are also subject to misuse and abuse. Because they may affect sexual risk, the ethics of their use has been questioned [37]. Our study addresses this question by examining whether sexual risk among HIV-positive men is related to testosterone use.


HIV-positive individuals are at increased risk of depression [38], and are often prescribed antidepressants by their providers. An analysis of the data from the HIV Cost and Services Utilization Study, a large probability sample of HIV-positive individuals, found that 27% of patients took psychotropic medicines in 1996, and antidepressants were the most commonly prescribed of these medicines (21% of patients) [39]. Among patients with mood disorders, the use of antidepressants was less common among African Americans (51%) than among Hispanic (67%) or Caucasian (61%) individuals [39].

Both depression and its treatment (particularly with serotonergic agents) can be related to ED [17,40]. Antidepressant medications that impair sexual functioning may further complicate safer sex efforts by men with ED. Men who avoid condoms because of difficulty in maintaining erections may find condoms even more problematical when antidepressants decrease sexual sensitivity. The links between steroid use and depression are complex, with steroid users being more likely to report suicidal thoughts in the previous 6 months (even when controlling for HIV status) and to have felt depressed [32]. The following links between depression and anabolic steroids have been found: (i) some studies show that depressed men have low testosterone levels; (ii) large doses of anabolic steroids can lead to mania or hypomania, and withdrawal from steroids is related to depressive symptoms; and (iii) men with low testosterone, including HIV-positive men, show symptoms of depression, and administering testosterone decreases depression [34,41]. Although the immediate effects of antidepressants may lead to the impairment of sexual functioning, over time these treatments should lead to better mood, increased interest in sex, and improved sexual functioning. To date, no study has examined the associations between the sexual practices among HIV-positive gay and bisexual men and antidepressant use.

The current study

The sexual functioning of HIV-positive men might be affected by their disease, HIV medications, or the other medications they take to treat problems such as depression or ED. For example, a depressed HIV-positive individual might be prescribed antiretroviral medications, antidepressants (for depression), viagra to help with ED (caused by a variety of factors), and testosterone (to decrease fatigue, increase lean body mass and improve libido). However, drugs that increase the ability to achieve and maintain an erection (such as viagra) or increase sexual desire or functioning over the long term (such as testosterone or antidepressants) may have an adverse public health effect on the AIDS epidemic by increasing the frequency of unprotected sex among HIV-positive gay and bisexual men and their partners of negative or unknown serostatus. This paper focuses on: (i) the prevalence of the use of viagra, testosterone, and antidepressants among HIV-positive gay and bisexual men; (ii) the bivariate correlates of the use of these substances, including sexual risk behavior; and (iii) the multivariate predictors of the use of these substances.

Materials and methods

Data were collected as part of the Seropositive Urban Men's Intervention Trial, a multisite randomized controlled trial of an HIV prevention intervention funded by the Centers for Disease Control and Prevention. Data reported in this study are from the baseline quantitative survey gathered using audio-computer assisted self-interviewing technology from a diverse sample of 1168 HIV-positive gay and bisexual men in New York City and San Francisco between March 2000 and June 2001. This study was approved by the Centers for Disease Control and Prevention Institutional Review Board (IRB) and by the IRB of participating institutions. The recruitment methods, eligibility, and demographics of the overall sample are described elsewhere [42].


Prescription drug use

After participants were asked about their use of antiretroviral medications, they were asked about their current use of 16 other medications commonly prescribed for HIV-positive men to treat a variety of conditions. We calculated the prevalence of the use of three substances that might affect sexual desire: viagra, testosterone, and antidepressants. We do not know whether the prescription drugs that men reported currently using were actually prescribed for them by their doctors.

Sociodemographic characteristics

The following demographic variables were collapsed into two to four categories for analyses: city, age, race/ethnicity, sexual orientation, educational background, and personal income.

Health status

Participants reported how many years they had been HIV positive, whether they had a diagnosis of AIDS, and whether or not they were taking antiretroviral medications.

Alcohol and illicit drug use

Participants were also asked about their drinking and their use of non-injection drugs in the past 90 days. Those who reported any drug use were asked to indicate whether they had used any of the following nine types or classes of drugs: speed/amphetamines, crystal (methamphetamine), barbiturates/tranquilizers, cocaine, ecstasy, special K (ketamine), marijuana (pot, hash), poppers (amyl nitrate), and gamma hydroxybutyrate (GHB). The number of these non-injection drugs used in the past 90 days (ranging from 0 to 9) was also calculated. Participants who drank alcohol were asked how many times in the past 90 days they drank ‘before or during sex’, and men who used non-injection drugs were asked how many times they had used drugs ‘before or during sex’. Finally, participants were asked whether they had ever injected drugs.

Psychological symptoms

Participants answered a portion of the Brief Symptoms Inventory (BSI) [43], specifically three subscales, depression (seven items, alpha 0.89), anxiety (six items, alpha 0.86), and hostility (five items, alpha 0.77).

Sexual risk behaviors

First we calculated the number of male sex partners in the past 3 months. Then, sexual behaviors were assessed by asking participants to indicate the frequency of four behaviors (insertive oral, receptive oral, insertive anal, receptive anal) in the 3 months before the completion of the survey. Participants reported how often they engaged in each sexual behavior with and without the use of condoms; separate frequencies were obtained for main partners (defined as ‘a partner you would call your boyfriend, spouse, significant other, or life partner’) and casual partners. Participants reported sexual behavior frequencies by serostatus (HIV positive, HIV negative, and unknown serostatus). Data for partners who were HIV negative or of unknown serostatus were combined (while still keeping main and non-main partners separate), and all frequencies were dichotomized. We were particularly interested in whether participants had engaged in unprotected insertive anal intercourse (UIAI), unprotected receptive anal intercourse (URAI), and unprotected insertive oral intercourse (UIOI) with different types of partners.


First we examined the prevalence of the use of viagra, testosterone, and antidepressants, and the prevalence of the use of substances that might interact dangerously with viagra. Then we conducted bivariate analyses to determine the correlates of the use of each of the drugs separately to help us understand differences for demographics, health status, alcohol and illicit drug use, psychological symptoms, and sexual risk between users and non-users of the three drugs. Chi-square analyses were used for categorical variables, and t-tests or analysis of variance were used for continuous variables, with the significance level set at P < 0.05. Finally, we entered all the significant bivariate variables into three separate multivariate logistic regressions to look at the predictors of the use of each of these three drugs. We were particularly interested in whether sexual risk behavior was associated with the use of any of these drugs.


Prescription drug use

Of the 1168 participants, 235 men (20.1%) reported using antidepressants, 225 (19.3%) used testosterone, and 144 (12.3%) used viagra. Overall, 314 men (26.9%) reported using one of these three drugs, 106 (9.1%) used two of these drugs, and 26 (2.2%) used all three drugs in the past 3 months. Of participants who were currently using viagra, 56 (38.9%) reported having used poppers in the past 3 months and 32 (22.2%) reported currently using ritonavir.

Bivariate associations between prescription drug use and other variables

Demographics and health status

The use of viagra was associated with being older, belonging to the white or ‘other’ racial group, self-identifying as gay, having more education, having an AIDS diagnosis, and taking antiretroviral medications (Table 1). The use of testosterone was significantly associated with being older, being white, being gay, having more education, having an AIDS diagnosis, and taking antiretroviral medications. Men who used testosterone had also been HIV positive, on average, for a longer time [9.9 versus 8.2 years, t(1149) = 4.59, P < 0.001]. The use of antidepressants was related to the city (more men in San Francisco took antidepressants), being older, belonging to the white or ‘other’ racial group, and having more education. Antidepressant use was not related to the three measures of health status.

Table 1:
Bivariate associations of viagra, testosterone, and antidepressants with demographic characteristics and health status.

Use of alcohol and illicit non-injection and injection substances

Drinking alcohol or using non-injection drugs in the past 3 months was associated with the use of viagra but not testosterone or antidepressants (Table 2). Participants who reported having ever injected illicit drugs were more likely to be taking antidepressants but not the other two drugs.

Table 2:
Bivariate associations of viagra, testosterone, and antidepressants with substance use.

We also examined whether the use of the three prescription drugs was associated with the number of illicit, non-injection drugs used in the past 3 months (range 0–9), or with the use of any of the nine specific illicit non-injection drugs. Participants who used viagra reported using more illicit substances in the past 3 months (M = 1.8) than men who did not use viagra (M = 1.2) [F(1, 1164) = 20.82, P < 0.001]. More than 47% of those who used viagra, and 33% of those who did not, had used two or more substances in the past 90 days. Men who were currently taking viagra were also more likely in the past 3 months to have taken ketamine (34.0 versus 14.1%, P < 0.001), and poppers (18.7 versus 13.2%, P = 0.05).

Men who were currently taking testosterone reported using more illicit substances in the past 3 months (M = 1.5) than men who were not (M = 1.2) [F(1, 1164) = 5.67, P = 0.02]. Approximately 43% of those who were taking testosterone, and 32% of those who were not, had used two or more substances in the past 3 months. Men who were currently taking testosterone were also more likely in the past 3 months to have taken barbiturates/tranquilizers (33.9 versus 20.2%, P = 0.02), and poppers (29.4 versus 14.8%, P < 0.001), and were less likely to have used cocaine (16.5 versus 23.5%, P = 0.04).

Participants who were currently taking antidepressants had used more illicit substances in the past 3 months (M = 1.6) than those who did not (M = 1.2) [F(1, 1164) = 12.39, P < 0.001]. Almost 46% of men taking antidepressants, and 32% of men who were not, had used two or more substances in the past 3 months. Men taking antidepressants were also more likely in the past 3 months to have taken barbiturates (35.7 versus 21.0%, P = 0.01), poppers (29.1 versus 16.7%, P = 0.001), and crystal (31.0 versus 20.4%, P = 0.01).

Drinking alcohol or using illicit non-injection drugs before or during sex

Illicit drug use before or during sex was associated with the use of all three prescription drugs, but alcohol use before or during sex was not related to the use of any of the three drugs. Viagra users reported more instances of mixing illicit substances with sex in the past 3 months (M = 11.7) than did participants who did not use viagra (M = 4.7) [F(1, 1152) = 16.92, P < 0.001]. Similarly, participants who used testosterone reported using illicit drugs before or during sex more times in the past 3 months (M = 9.7) than men who did not use testosterone (M = 4.6) [F(1, 1152) = 13.01, P < 0.001], as did those who took antidepressants (M = 7.9) compared with those who did not (M = 5.0) [F(1, 1152) = 4.23, P < 0.04].

Psychological symptoms

Participants who were taking viagra had significantly higher hostility scores on the BSI than men not taking viagra, and participants who were taking testosterone had significantly higher depression scores on the BSI than men not taking testosterone (data not shown). Participants who were taking antidepressants had significantly higher scores on the BSI for depression, anxiety, and hostility (data not shown).

Sexual behavior

First, we examined the total number of male partners in the past 3 months for those who did and did not use prescription drugs, and found differences for all three drugs. Participants who used viagra had more partners (M = 15.9) than those who did not (M = 8.2) [F(1, 1162) = 16.70, P < 0.001], men who took testosterone had more partners (M = 12.2) than those who did not (M = 8.5) [F(1, 1162) = 5.63, P = 0.02], and men taking antidepressants had more partners (M = 13.2) than those who did not (M = 8.2) [F(1, 1162) = 10.67, P = 0.001].

Then we conducted six sets of chi-square analyses for each of the three prescription drugs, and examined various sexual risk behaviors by partner serostatus for both casual partners (Table 3) and main partners. For casual partners, more participants who used viagra engaged in all three risk behaviors (UIAI, URAI, and UIOI) with both types of partners (HIV positive and HIV negative/unknown serostatus) than did those who did not use viagra. Significantly more of those who took testosterone or antidepressants engaged in URAI and UIOI with casual partners who were HIV positive and who were HIV negative/unknown status than did those who did not take testosterone or antidepressants. However, men who took testosterone or antidepressants were not more likely to engage in UIAI with non-main partners. For main partners, there was only one significant finding, men who used testosterone were more likely to engage in URAI with their HIV-positive main partners (10.7%) compared with men who did not use testosterone (6.1%) (P = 0.02). Their data were in the same direction for UIAI with HIV-positive main partners (10.7 versus 7.0%, NS). Because most of the significant associations between the use of these three prescription drugs and sexual risk were for non-main partners, multivariate analyses included only the univariate findings with non-main partners.

Table 3:
Sexual risk behavior with non-main partners by partner serostatus for users and non-users of viagra, testosterone, and antidepressants.

Multivariate associations

We performed logistic regressions to determine which variables were predictive of the use of each drug, using variables that were significant in bivariate analyses (Table 4). The predictors of viagra use were being older, being more educated, using ketamine in the past 3 months, and engaging in UIOI with partners who were HIV negative/unknown seroserostatus. There was a trend for UIAI with HIV-positive casual partners to be associated with viagra use (P = 0.06). The predictors of testosterone use were having more education, having AIDS, and using poppers in the past 3 months. The predictors of antidepressant use were race (Hispanic individuals were less likely to be on antidepressants), high depression scores on the BSI, and using poppers in the past 3 months. There was a trend for more crystal use to predict the use of antidepressants (P = 0.06).

Table 4:
Logistic regression models showing significant predictors of use of viagra, testosterone, and antidepressants.


Prescription medications used by HIV-positive men can be used appropriately to treat valid medical conditions, but they can also be misused or have unintended negative effects, such as medication interactions or associations with risky sexual behavior. This is particularly the case for a drug such as viagra, which is fast acting, short lasting, and provides a very desirable effect. We found that 20% of participants were currently using antidepressants, 19% were using testosterone, and 12% were using viagra. The proportion of men using antidepressants was very similar to the findings from the HIV Cost and Services Utilization Study [39]. The proportion of men ‘currently’ using viagra in our study was lower than other studies conducted during the same time period [22,23], possibly because of longer reporting periods or different sampling methods in those studies. Our data also suggest that men may be mixing viagra in potentially dangerous combinations, particularly with ritonavir and poppers. Knowing more about the timing of use and strategies used to avoid potentially dangerous drug interactions would help us understand more about men who use various prescription and non-prescription drugs. Unfortunately, we could not assess whether viagra was taken at the same time as poppers, making it difficult to assess the actual risk. However, some men did report taking ritonavir and viagra currently, although we could not assess whether they took smaller doses of viagra or used any other strategy suggested by their physician to decrease the likelihood of drug interactions. Clearly, care providers should discuss drug interactions when they prescribe either ritonavir (as some men will obtain viagra without a prescription) or viagra, and they should work with HIV-positive patients to understand the potential dangers.

Interesting patterns emerged for the bivariate correlates of use of these three prescription drugs. We found evidence of the appropriate use of these three medicines to treat existing medical needs. For example, men who were depressed were more likely to be taking antidepressants and testosterone, both of which may help with depression and well being. Similarly, although we did not measure ED or testosterone levels, men who had AIDS or were taking antiretroviral medications were more likely to be taking viagra and testosterone, and men who had been HIV positive longer were more likely to be taking testosterone. Either of these drugs would be more indicated, on average, for men who are sicker or taking other medications.

Bivariate analyses of drinking alcohol and using illicit substances showed that only the use of viagra was related to the use of alcohol and the use of illicit non-injection drugs in the past 3 months. This finding is consistent with previous research showing an association between viagra use and illicit substance use [20,21–24], and expands the literature to include alcohol. This may be because the fast-acting, short-term nature of viagra is more similar to the action of alcohol and illicit drugs, whereas the effects of testosterone and antidepressants occur over time and more gradually. Having ever injected illicit drugs was related to depression, which supports the link between substance abuse or addiction and depression [38]. A few of the nine specific drugs were related to the use of these three prescription drugs, but using poppers was the only single-drug variable that was significantly related to the use of all three prescription drugs at the bivariate level.

Regarding sexual risk, men who used viagra, but not those who used testosterone or antidepressants, were more likely to engage in UIAI with both their HIV-negative/unknown-status casual partners and their HIV-positive partners. The former sexual behavior is the riskiest in terms of the likelihood of transmitting HIV to an uninfected partner. Similar to previous research, there was some evidence that HIV-positive men who use viagra were most likely to engage in risk with their positive partners than their HIV-negative/unknown-status partners [21,23], but our data also show an increased risk to the HIV-negative/unknown-status partners of the viagra using men. Given the physiological action of viagra, it is not surprising that it was associated with UIAI, although we cannot determine if men who wanted to engage in this behavior were more likely to be drawn to the drug or whether using the drug somehow led to men engaging in this behavior. The use of all three prescription drugs was strongly associated with URAI and UIOI with both HIV-positive casual partners as well as with HIV-negative/unknown-serostatus casual partners and was associated with URAI with HIV-positive main partners. Although receptive anal sex and insertive oral sex behaviors are lower risk than insertive anal sex, these behaviors still carry varying risks to all partners. Clearly, HIV-positive users of each of these three drugs were more likely to put their partners at risk of HIV, other sexually transmitted infections, or potential superinfection with HIV, and this was particularly the case with viagra. Finally, it was interesting that the only effect for main partners was more URAI among testosterone users. Because testosterone works over time to increase energy and libido, men with HIV-positive main partners may experience increased sexuality over time, which includes being more willing to have unprotected sex in an intimate relationship. The data were in the same direction for UIAI with HIV-positive main partners, which supports this explanation.

In the multivariate analysis, the predictors of viagra use were being older, being more educated, using ketamine in the past 3 months, and engaging in UIOI with casual partners who were HIV negative or of unknown serostatus. A trend showed increased UIAI with HIV-positive partners. Men who engage in more UIAI with their HIV-positive partners and more UIOI with their partners who are HIV negative or of unknown serostatus are thus more likely to take viagra. These multivariate findings do not diminish the importance of the univariate findings, because some of the significant variables in the multivariate model, such as the use of ketamine, may be important proxies for sexual risk.

Multivariate predictors of testosterone use were having more education, having AIDS, and having used poppers in the past 3 months. Multivariate predictors of antidepressant use were not being Hispanic, high depression scores on the BSI, and using poppers. There was also a trend for more use of crystal (methamphetamines) in the past 3 months among viagra users. Given the different modes of action of these two substances and viagra, it is not surprising that they were not related to unprotected sex in the multivariate model. Popper use was also related to the use of testosterone and antidepressants. Given that poppers are used primarily during sex, it may be that men taking testosterone and antidepressants are having improved sexual lives and are using poppers to enhance their pleasure. Given the relationship between poppers and risk for HIV-positive men [9], HIV prevention efforts should include those men who use poppers. The finding for methamphetamine use being related to taking antidepressants is interesting because depression is a common side-effect shortly after coming down from a binge [44], and abnormalities in those parts of the brain that are implicated in mood disorders have been found among methamphetamine abusers [45]. We cannot determine, however, whether men who use methamphetamines may take antidepressants more to treat depressed feelings or whether depression may precede the use of the illicit drug. In either instance, it is important for physicians to educate their patients about the side-effects of drugs such as methamphetamine and explain the futility of treating depression while taking illicit drugs that contribute to depression.

Finally, it is interesting that demographics were predictive of the use of these prescription drugs in multivariate analyses. Viagra use was more common among older and more educated men, testosterone use was more common among more educated men, and antidepressant use was less common among Hispanic men. A lack of education and race may thus be a barrier to getting or being willing to take these potentially beneficial medications. This suggests that physicians must continue actively to advocate pharmacological treatments for their patients who are less able to do so for themselves as a result of stigma, shame, cultural pressures, marginalization, or other internal or external forces. Reassuringly, having AIDS was related to taking testosterone, suggesting that testosterone is not being widely abused by men with less advanced HIV disease. Similarly, being more depressed was related to the current use of antidepressants.

The overall limitations of the Seropositive Urban Men's Intervention Trial are discussed elsewhere [42], but there are a few limitations of these particular analyses. First, our data were collected for the ‘current’ use of these three drugs. It is thus difficult to compare our data with other studies asking for longer time periods. Therefore, it is unclear whether the same relationships would emerge among a broader cross-section of men who might have used viagra or the other two drugs in some past time period but are not currently using it. Second, we cannot determine whether the men used the various prescription drugs at the same time as illicit substances or whether the medications were prescribed for them or obtained in another manner. A final limitation is that the data are cross-sectional, so there is no way to imply causation. We cannot thus know whether viagra use leads to unprotected sex, whether such sex is planned in advance and viagra helps it happen, or whether some third variable such as having a risk-taking personality explains both substance use and transmission risk.

These results have important implications for public health. First, HIV-positive MSM reported moderate rates of viagra use and the frequent use of substances that can potentially interact dangerously with viagra. In addition, men who used viagra drugs were more likely to engage in various risk behaviors with partners of every serostatus. In bivariate analysis, but not multivariate analysis, the use of testosterone or antidepressants was associated with sexual risk. In multivariate analysis, however, the use of testosterone or antidepressants was related to the use of poppers, which may be a proxy measure of sexual risk. Medical providers who prescribe viagra to HIV-positive patients need to assess the patients’ risk, and counsel them about potential medication interactions and the need for protection from sexual risks [46]. Similarly, providers who prescribe testosterone and antidepressants should counsel their patients about sexual risk, even though these risks may emerge more slowly as the drugs take effect over time. Future research should examine the nexus between sexual risk and taking both prescription and illicit substances. Through a further understanding of these interactions, care providers can better ensure that their HIV-positive patients get the most benefit out of these prescription drugs while reducing the potential public health harms that come from their use and abuse.


The authors would also like to acknowledge the following people who contributed to this research: Cynthia Gómez, Robert Hays, Ann O'Leary, Robert Remien, David Bimbi, Tim Matheson, Byron Mason, Carmen Mandic, Bonnie Faigeles, Nick Alvarado, Andrew Nelson Peterson, Eric Rodriguez, Paul Galatowitsch, Michael Marino, Aongus Burke, Michael Stirratt, Eric Martin, Gloria Abitol, Caroline Bailey, and Cindy Lyles.

Sponsorship: Research on SUMIT was funded by the Centers for Disease Control and Prevention through cooperative agreements with New Jersey City University (UR3/CCU216471) and the University of California, San Francisco (UR3/CCU916470).


1. Centers for Disease Control and Prevention. Primary and secondary syphilis among men who have sex with men – New York City, 2001.MMWR 2002; 51:853–856.
2. Centers for Disease Control and Prevention. Increases in HIV diagnoses – 29 states, 1999–2002. MMWR 2003; 52:1145–1148.
3. Chen SY, Gibson S, Weide D, McFarland W. Unprotected anal intercourse between potentially HIV-serodiscordant men who have sex with men in San Francisco. J Acquir Immune Defic Syndr 2003; 33:166–170.
4. Wolitski RJ, Valdiserri RO, Denning PH, Levine WC. Are we headed for a resurgence in the HIV epidemic among men who have sex with men? Am J Public Health 2001; 91:883–888.
5. Chesney MA, Barrett DC, Stall R. Histories of substance use and risk behavior precursors to seroconversion in homosexual men. Am J Public Health 1998; 88:113–116.
6. Stall R, Purcell DW. Intertwining epidemics: a review of research on substance use among men who have sex with men and its connection to the AIDS epidemic. AIDS Behav 2000; 4:181–192.
7. Klitzman RL, Pope HG, Hudson JI. MDMA (“ecstasy”) abuse and high-risk sexual behaviors among 169 gay and bisexual men. Am J Psychiatry 2000; 157:1162–1164.
8. McNall M, Remafedi G. Relationship of amphetamine and other substance use to unprotected intercourse among young men who have sex with men. Arch Pediatr Adolesc Med 1999; 153:1130–1135.
9. Purcell DW, Parsons JT, Halkitis PN, Mizuno Y, Woods WJ. Substance use and sexual transmission risk behavior of HIV-positive men who have sex with men. J Subst Abuse 2001; 13:185–200.
10. Nehra A, Kulaksizoglu H. Global perspectives and controversies in the epidemiology of male erectile dysfunction. Curr Opin Urol 2002; 12:493–496.
11. Merry C, Barry MG, Ryan M, Tjia JF, Hennessy M, Eagling VA, et al. Interaction of sildenafil and indinavir when co-administered to HIV-positive patients. AIDS 1999; 13:F101–F107.
12. Simons JS, Carey MP. Prevalence of sexual dysfunctions: results from a decade of research. Arch Sex Behav 2001; 30:177–219.
13. Tindall B, Forde S, Goldstein D, Ross MW, Cooper DA. Sexual dysfunction in advanced HIV disease. AIDS Care 1994; 6:105–107.
14. Colson AE, Keller MJ, Sax PE, Pettus PT, Platt R, Choo PW. Male sexual dysfunction associated with antiretroviral therapy. J Acquir Immune Defic Syndr 2001; 30:27–32.
15. Collazos J, Mayo J, Maretinez E, Ibarra S. Association between sexual disturbances and sexual hormones with specific antiretroviral drugs. AIDS 2002; 16:1294–1295.
16. Solima S, Osio M, Muscia F, Gambaro P, Alciati A, Zucconi M, et al. Protease inhibitors and erectile dysfunction [Letter]. AIDS 2001; 15:2331–2333.
17. Physician's Desk Reference, 58th ed. Montvale, NJ: Medical Economics Company; 2004. Available from http://www/
18. Stall R, Paul JP, Greenwood G, Pollack LM, Bein E, Crosby GM, et al. Alcohol use, drug use and alcohol-related problems among men who had sex with men: the Urban Men's Health Study. Addiction 2001; 96:1589–1601.
19. Aldridge J, Measham F. Sildenafil (Viagra) is used as a recreational drug in England [Letter]. BMJ 1999; 318:669.
20. Colfax GN, Mansergh G, Guzman R, Vittinghoff E, Marks G, Rader M, Buchbinder S. Drug use and sexual risk behavior among gay and bisexual men who attend circuit parties: a venue-based comparison. J Acquir Immune Defic Syndr 2001; 28:373–379.
21. Sherr L, Bolding G, Maguire M, Elford J. Viagra use and sexual risk behaviour among gay men in London [Letter]. AIDS 2000; 14:2051–2053.
22. Chu PL, McFarland W, Gibson S, Weide D, Henne J, Miller P, et al. Viagra use in a community-recruited sample of men who have sex with men, San Francisco. J Acquir Immune Defic Syndr 2003; 33:191–193.
23. Kim AA, Kent CK, Klausner JD. Increased risk of HIV and sexually transmitted disease transmission among gay or bisexual men who use Viagra, San Francisco 2000–2001. AIDS 2002; 16:1425–1428.
24. Halkitis PN, Parsons JT, Wilton L. An exploratory study of contextual and situational factors related to methamphetamine use among gay and bisexual men in New York City. J Drug Issues 2003; 33:413–432.
25. Mooney M, Vergel NR. Built to survive: a comprehensive guide to the medical use of anabolic steroids, nutrition, and exercise for HIV (+) men and women. Houston, TX: Program for Wellness Restoration; 2003. Available at:
26. Newshan G, Leon W. The use of anabolic agents in HIV disease. Int J STD AIDS 2001; 12:141–144.
27. Wagner G, Rabkin J, Rabkin R. Effects of testosterone replacement therapy on sexual interest, function and behavior in HIV+ men. J Sex Res 1997; 34:27–33.
28. Halkitis PN. An exploration of perceptions of masculinity among gay men living with HIV. J Men's Studies 2001; 9:413–429.
29. Leit RA, Pope HG Jr, Gray JJ. Cultural expectations of muscularity in men: the evolution of playgirl centerfolds. Intl J Eating Disord 2001; 29:90–93.
30. Bolding G, Sherr L, Maguire M, Elford J. HIV risk behaviours among gay men who abuse anabolic steroids. Addiction 1999; 94:1829–1835.
31. Elford J, Bolding G, Maguire M, Sherr L. Do gay men discuss HIV risk reduction with their GP? AIDS Care 2000; 12:287–290.
32. Bolding G, Sherr L, Elford J. Use of anabolic steroids and associated health risks among gay men attending London gyms. Addiction 2002; 97:195–203.
33. Halkitis PN, Green KA, Wilton L. Masculinity, body image, and sexual behavior in HIV-seropositive gay men: a two-phase formative behavioral investigation using the internet. Intl J Men's Health 2004; 3:27–42.
34. Pope HG, Cohane GH, Kanayama G, Siegal AJ, Hudson JI. Testosterone gel supplementation for men with refractory depression: a randomized, placebo-controlled trial. Am J Psychiaty 2003; 160:105–111.
35. Middleman AB, Faulkner AH, Woods ER, Emans SJ, DuRant RH. High-risk behaviors among high school students in Massachusetts who use anabolic steroids. Am Acad Pediatrics 1995; 96:268–272.
36. Midgley SJ, Heather N, Best D, Henderson D, McCarthy S, Davies JB. Risk behaviours for HIV and hepatitis infection among anabolic-androgenic steroid users. AIDS Care 2000; 12:163–170.
37. Fisher WA. Do no harm: on the ethics of testosterone replacement therapy for HIV+ persons. J Sex Res 1997; 34:35–36.
38. Cruess DG, Evans DL, Repetto MJ, Gettes D, Douglas SD, Peitto JM. Prevalence, diagnosis, and pharmacological treatment of mood disorders in HIV disease. Biol Psychiatry 2003; 54:307–316.
39. Vitiello B, Burnam MA, Bing EG, Beckman R, Shapiro MF. Use of psychotropic medications among HIV-infected patients in the United States. Am J Psychiatry 2003; 160:547–554.
40. American Psychiatric Association. Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: American Psychiatric Association; 1994.
41. Rabkin JG, Wagner GJ, Rabkin R. A double-blind placebo-controlled trial of testosterone therapy for HIV-positive men with hypogonadal symptoms. Arch Gen Psychiatry 2000; 57:141–147.
42. Wolitski RJ, Parsons JT, Gómez CA, Purcell DW, Hoff CC, Halkitis PN. Prevention with gay and bisexual men living with HIV: rationale and methods of the Seropositive Urban Men's Intervention Trial. AIDS 2005; 19(suppl. 1):S1–S11.
43. Derogatis LR, Melisaratos N. The brief symptom inventory: an introductory report. Psychol Med 1983; 13:595–605.
44. Freese TE, Miotto K, Reback CJ. The effects and consequences of selected club drugs. J Subst Abuse Treatment 2002; 23:151–156.
45. London ED, Simon SL, Berman SM, et al. Mood disorders and regional cerebral metabolic abnormalities in recently abstinent methamphetamine abusers. Arch Gen Psychiatry 2004; 61:73–84.
46. Centers for Disease Control and Prevention, Health Resources and Services Administration, National Institute of Health, HIV Medicine Association of the Infections Disease Society of America, HIV Prevention in Clinical Care Working Group. Recommendations for incorporating human immunodeficiency virus (HIV) prevention into the medical care of persons living with HIV. Clin Infect Dis 2004; 38:104–121.

antidepressants; gay and bisexual men; HIV-positive; men who have sex with men; sexual risk behavior; testosterone; viagra

© 2005 Lippincott Williams & Wilkins, Inc.