| |
Human Immunodeficiency Virus in Semen and Plasma: Investigation of Sexual Transmission
Risk Behavioral Correlates
|
| |
| |
Click here to download PDF
.....Type of sexual behavior, unprotected insertive sex, and STIs associated with HIV levels in semen....
SETH C. KALICHMAN,1 MARJORIE CAGE,1 TAMARA BARNETT,2 PHILLIP THARNISH,2 DAVID ROMPA,1 JAMES AUSTIN,1 WEBSTER LUKE,1
JULIE O'MOWREY,2 and RAYMOND F. SCHINAZI2
1Center for AIDS Intervention Research, Medical College of Wisconsin, Milwaukee, Wisconsin 53226.
2Emory University School of Medicine and Veterans Affairs Medical Center, Decatur, Georgia 30033.
AIDS RESEARCH AND HUMAN RETROVIRUSES
Volume 17, Number 18, 2001, pp. 1695-1703
We found that 53% of men with undetectable plasma viral loads had measurable virus in their semen
the majority of HIV transmission risk behavior in our sample occurred
among men who had the highest concentration of HIV in their semen. It is noteworthy that differences in sexual activity between men with higher and relatively lower concentrations of HIV in semen were significant only for insertive sexual acts. One factor that may explain these findings was the potential co-occurrence of other STIs.
The association of higher rates of insertive intercourse and semen viral loads may therefore be a function of asymptomatic or subclinical urethritis
We recommend that future research investigating relationships between sexual transmission risk behavior and semen viral loads perform more comprehensive
clinical examinations for clinical and subclinical urethritis and conduct more comprehensive testing for STIs on larger samples. We also recommend that HIV-positive men be warned that they cannot infer their infectiousness from their
plasma viral load test results and that behavioral risk reduction practices remain the only means for preventing the spread of HIV.
Analyses of sexual transmission risk behaviors showed that men with greater concentrations of HIV in their semen relative to concentrations of HIV in plasma reported significantly greater rates of unprotected vaginal intercourse and total number of unprotected sexual intercourse occasions as the insertive partner than men with equal to or greater concentrations of virus in their plasma than semen.
We were particularly interested in patterns of transmission risk behaviors
exhibited by men who were more infectious than implicated by their plasma viral load. Given that greater concentrations of virus are typically detected in plasma relative to serum,12 and the demonstrated association between urethritis and viral load in semen,17 we hypothesized that men who presented higher viral
loads in semen than blood would be more likely to have a recent STI and would exhibit higher rates of HIV transmission risk behaviors than men whose plasma viral load was equal to or less than their semen viral load.
ABSTRACT
Risks for sexually transmitted HIV may be related to concentrations of virus detected in semen and previous research shows a small to moderate association between viral load in blood and semen. This study examined the association between viral load in semen and plasma in a community sample of HIV-infected men and is the first study to examine semen viral load in relation to sexual transmission risk behaviors. A sample of 44 HIV-positive men recruited from community service agencies provided semen, blood, and urine samples and completed clinical interviews assessing health and behavior. We failed to find an association between viral load in semen and plasma, Spearman p=0.07, p >0.1. When restricted to participants with detectable virus in semen and plasma, the correlation remained nonsignificant, p= -0.16, p >0.1. Men who had higher semen
viral loads relative to their plasma viral load were distinguished by having engaged in significantly higher rates of unprotected intercourse as the insertive sex partner in the previous 3 months. Semen viral load was not, however, related to recent or current sexually transmitted infections (STIs). This study is among the first to examine sexual transmission risk behaviors as marker for HIV infectiousness. Results caution against inferring sexual transmission infectiousness based on plasma viral load and suggest that HIV-positive men who practice higher rates of insertive intercourse may be more infectious even in the absence of other STIs.
INTRODUCTION
STUDIES SHOW that approximately one in three people living with HIV AIDS continues to practice unprotected sexual behaviors, potentially placing themselves and their sex partners at risk for exposure to HIV and other sexually transmitted pathogens. 1,2 Among the factors associated with continued sexual
transmission risk behaviors in HIV-infected people are the perceived benefits of medical advances in HIV treatments.3,4 Successful antiretroviral therapy suppresses HIV replication and can reduce the amount of HIV in peripheral blood to levels below detection, suggesting reduced infectivity. Thus, individuals
with undetectable virus in peripheral blood may perceive themselves as less infectious and perceptions of reduced infectivity appear to influence sexual transmission risk behaviors in some HIV-infected and uninfected persons.3,5-7 There is also evidence that individuals with higher concentrations of plasma HIV are more likely to transmit the virus to blood recipients, offspring, and sex partners.8-10 However, studies that correlate plasma levels of HIV RNA with viral RNA detected in semen demonstrate low to moderate associations; most correlations range between 0.20 and 0.60.9-12 In general, plasma concentrations of HIV tend to be higher than concentrations in semen, 12,13 with HIV RNA as much as 10-fold higher in plasma than semen.11 Accumulating evidence suggests that the male genital tract constitutes a relatively distinct compartment of HIV replication, with long-lived and frequently replicating cells serving as a sanctuary for HIV to replicate protected from immune responses and adequate levels of some antiretroviral therapies. 8 Explanations of discrepancies between HIV viral loads in plasma and semen have thus far focused on co-occurring ulcerative and nonulcerative sexually transmitted infections (STIs) as factors that promote HIV shedding in the genital tract and therefore potentially increasing infectivity in people living with HIV AIDS.14 In one study, HIV concentrations in semen were 8 times greater in men with clinical urethritis and 10 times greater in men with gonorrhea relative to those without clinical urethritis.15 In addition, successful treatments of STIs demonstrate subsequent reductions of HIV in semen,16 and treatment of symptomatic STIs can reduce HIV transmission to uninfected sex partners.17 Thus, co-occurring STIs appear important in promoting infectivity in HIV-infected men. In addition to infectiousness, HIV transmission is a function of sexual transmission risk practices. Studies have not as of yet, however, investigated behavioral cofactors of infectivity in HIV-infected men.
The current study was conducted to identify behavioral correlates of viral infectivity in HIV-positive men. We were particularly interested in patterns of transmission risk behaviors exhibited by men who were more infectious than implicated by their plasma viral load. Given that greater concentrations of virus are typically detected in plasma relative to serum,12 and the demonstrated association between urethritis and viral load in semen,17 we hypothesized that men who presented higher viral loads in semen than blood would be more likely to have a recent STI and would exhibit higher rates of HIV transmission risk behaviors than men whose plasma viral load was equal to or less than their semen viral load.
DISCUSSION
In contrast to previous research reporting moderate degrees of association between HIV concentrations in plasma and semen, the current study found no such relationship. We also found poor concordance between undetectable levels of HIV RNA in plasma and semen. A factor that may account for these discrepancies is differences in samples under study. Unlike previous investigations of HIV levels detected in semen, the current sample was not solely recruited from infectious disease clinics; one in five of our participants were not currently receiving antiretroviral medications. In addition, our participants
who were receiving treatment represented a wide range of antiretroviral
regimens, with potential differences among drugs in their penetration of semen.23 Therefore, our data suggest that degrees of sexual infectivity in a heterogeneous sample of HIVinfected men cannot be estimated from plasma viral loads. We found that 34% of HIV-infected men had greater concentrations
of HIV in their semen than their plasma-the inverse of the expected
direction of association, and a proportion that is similar to the 38% of men with greater semen than plasma viral loads reported by Tachet et al.12 Therefore, a considerable number of HIV-infected men are more infectious than they may assume on the basis of results of viral load assayed from peripheral
blood. Although the threshold for HIV infectivity is not yet established
and persons with viral loads below levels of detection are likely infectious, studies show that HIV-positive and HIV-negative persons perceive less risk for contracting HIV when HIV-infected persons have undetectable plasma viral loads.3,6,7 Transmission risk perceptions may therefore be particularly
hazardous because of the inability to infer sexual infectivity from plasma viral loads.
The lack of association between viral load in plasma and semen in the current study could not be accounted for by HIV disease status, antiretroviral therapy history, or current treatment adherence. Some antiretrovirals do not penetrate semen as effectively as they do blood, allowing semen to harbor replication-
competent virus that can be sexually transmitted.13,23,24 We found that 53% of men with undetectable plasma viral loads had measurable virus in their semen. In addition, the majority of HIV transmission risk behavior in our sample occurred
among men who had the highest concentration of HIV in their semen. It is noteworthy that differences in sexual activity between men with higher and relatively lower concentrations of HIV in semen were significant only for insertive sexual acts. One factor that may explain these findings was the potential
co-occurrence of other STIs. However, we failed to identify current Gonorrhea, Chlamydia, or clinically reported recent STIs in greater proportions among men with higher viral loads in their semen than plasma. However, our ability to detect associations between semen viral load and STIs was limited by our relatively small sample size. The association of higher rates of insertive intercourse and semen viral loads may therefore be a function of asymptomatic or subclinical urethritis. We are not aware of research that has tested for subclinical urethritis in relation to HIV viral load in semen. We recommend that future research investigating relationships between sexual transmission risk behavior and semen viral loads perform more comprehensive clinical examinations for clinical and subclinical urethritis and conduct more comprehensive testing for STIs on
larger samples. We also recommend that HIV-positive men be warned that they cannot infer their infectiousness from their plasma viral load test results and that behavioral risk reduction practices remain the only means for preventing the spread of HIV.
RESULTS
Among the 242 men who were asked if they would be interested in participating in a study that involved semen collection, 54 declined. Analyses were conducted to identify potential volunteer biases by comparing three groups: (1) men who
declined interest in participating (n =54), (2) men who indicated an interest in participating but were not asked to enroll (n =138), and men who were randomly selected, approached, and enrolled in the study (n =50). Analyses did not indicate any differences in age, years of education, ethnicity, sexual orientation,
current relationship status, sexual behaviors, years since testing HIV positive, most recent CD41 cell count, and most recent viral load. We therefore failed to identify potential selection and volunteer biases along key participant characteristics.
Associations between viral load in plasma and semen
The Spearman's r correlation between viral load in semen and plasma was not significant, p (44) =0.07, p >0.1 (see Fig. 1). When restricted to participants with detectable virus in semen and plasma, the correlation remained nonsignificant, p (20) = -0.16, p >0.1 (see Fig. 2). There was also poor concordance between undetectable viral loads in plasma and semen; 53% of men with undetectable virus in plasma had detectable viral loads in semen and 31% of men with undetectable virus in semen had detectable plasma viral loads, x2 (1, n =
44) = 1.0, p >0.1. The contingency coefficient for undetectable viral load in blood and semen was also nonsignificant, CC =0.15, p >0.1.
Demographic and health status characteristics
For all participants, the median absolute difference between log values of viral load in semen and plasma was 0.7, with greater values in plasma than semen. The greatest difference between semen and plasma values was 2.7 log copies. A total of 15 (34%) men had greater HIV concentrations in semen than plasma. Among men with greater viral loads in their plasma than semen, a mean difference of 1.9 log copies of HIV was found in plasma relative to semen. For men who had greater viral loads in their semen than plasma, the mean difference was 3.5 log copies of HIV in semen relative to plasma. Table 1 presents the demographic and health status characteristics of men with equal or greater viral loads in their plasma than semen in comparison with men with greater concentrations of HIV in their semen than plasma. Men with equal or greater
virus in their plasma than their semen compared with men with greater virus in their semen than plasma showed no differences in age, ethnicity, sexual orientation, years since testing HIV positive, HIV symptoms experienced, CD41 cell counts, current HIV treatment status, or current HIV treatment adherence.
Sexual practices and sexually transmitted infections
Table 2 presents the individual participant log values for viral load in plasma and semen, sorted in ascending order of viral load values in semen, as well as frequencies of insertive sexual intercourse and current antiretroviral regimens. Of 44 men in the study, 16 had semen viral load lower than the level of detection and 15 had undetectable virus in their plasma; among men with detectable viral loads, 7 had both undetectable viral loads in semen and plasma. We grouped together the 15 (34%) men who had viral loads in semen that were higher than viral loads in plasma. Analyses of sexual transmission risk behaviors showed that men with greater concentrations of HIV in their semen relative to concentrations of HIV in plasma reported significantly greater rates of unprotected vaginal intercourse and total number of unprotected sexual intercourse occasions
as the insertive partner than men with equal to or greater concentrations of virus in their plasma than semen. The difference for unprotected insertive anal intercourse approached significance (p <0.06). Differences between groups for receptive intercourse acts and numbers of sex partners were not significant (see Table 3).
As shown in boldface in Table 2, five participants had been
diagnosed with a sexually transmitted infection in the previous
3 months (participants 5, 21, 25, 26, and 37) and one participant
was diagnosed with chlamydia via urine specimens (participant
20). Results failed to find differences in STI occurrences
between men with relatively higher and lower semen
viral loads.
|
|
| |
| |
|
|
|
