You are leaving Medscape Education
Cancel Continue
Log in to save activities Your saved activities will show here so that you can easily access them whenever you're ready. Log in here CME & Education Log in to keep track of your credits.


Sexual Transmission of HIV-1: New Data from the 10th CROI

Authors: Stephen Taylor, MD, MRCP, PhDFaculty and Disclosures



Sexual transmission of HIV continues to be the number-one mechanism for the sustained spread of the HIV epidemic in most countries throughout the world. Hence, this remains a key area for HIV research. For enthusiasts of the subject, the 10th Conference on Retroviruses and Opportunistic Infections in Boston, Massachusetts, contained a plethora of related abstracts that clarified existing theories or postulated new ideas pertaining to the sexual transmission of HIV. This short overview aims to highlight some of the more clinically relevant presentations pertaining to this fascinating topic.

Increasing Rates of HIV and Syphilis in the United States

In a plenary lecture, Dr. Valdiserri[1] from the U.S. Centers for Disease Control and Prevention (CDC) in Atlanta, Georgia, reviewed the data pertaining to the current state of the US epidemic. It is currently estimated that between 850,000 and 950,000 persons in America are living with HIV or AIDS. Data show that HIV incidence was stable throughout the 1990s at approximately 40,000 new infections per year -- down from an estimated high of 150,000 annual infections in the mid-1980s; however, Valdiserri highlighted a worrisome recent upward trend in HIV incidence.

In addition, recent HIV figures and national syphilis rates have led to concerns about a possible upturn in HIV incidence within the United States. Concentrating on the recent syphilis outbreaks involving men who have sex with men in cities such as Miami, New York City, and Los Angeles, Valdiserri revealed that investigations of these outbreaks showed that a worrying 43% to 59% of the men with syphilis were already known to be HIV-positive, with many already under care for their HIV disease.

The concern here is that individuals who are able to transmit or acquire syphilis are also engaging in activities capable of transmitting or acquiring HIV. Clearly, these data support the need for routine sexual health checks for HIV-positive persons, both for their own health and for the potential of reducing the spread of HIV and other sexually transmitted infections (STIs).

With this resurgence in HIV and STD incidence, several groups investigated novel areas for intervention, as discussed below.

The Role of the Internet for Sexual Transmission and Intervention

In the 1970s and 1980s, bathhouses and back rooms were associated with high-risk sexual behavior. At this meeting, Chaisson and colleagues[2] from John Hopkins University in Baltimore, Maryland, postulated that the Internet was the new "venue" to meet and arrange sexual liaisons. Furthermore, they found a significant association between meeting a sex partner online and having unprotected anal sex (UAS), particularly for HIV-positive men.

In their online study, they recruited 2934 men through internet chat room banners on a popular gay Web site. Respondents completed an anonymous, 60-item online questionnaire about sexual behavior and drug and alcohol use during a recent 6-month period. The HIV-negative group included those testing negative and those untested. Most (80%) reported sex only with men, 19% with men and women, and < 1% only with women. Forty-six percent were 18-29 years of age, 46% were 30-49, and 8% were 50 or older. Most men were white (85%) and had at least some college education (87%); 7.6% were HIV-positive. More than 100 lifetime sex partners were reported by 27%, with 6% reporting more than 10 partners in the past 30 days. Most men (84%) met sex partners online, and they were significantly more likely to have UAS than those who met partners in other ways (64% vs 58%, P = .02, chi-square test). Among those who met partners online, HIV-positive men were more likely to report UAS than were HIV-negative men (OR 1.5, P = .03).

Of the 10 men diagnosed with syphilis during the 6-month period, 9 of these met their sex partners online and 4 reported being HIV-positive.

This study confirms what many clinicians might already be seeing in their day-to-day clinical practice. On a more positive note, the ability to recruit patients via the Internet might provide a unique opportunity to use this medium as a means of providing urgently needed safer-sex messages.

HIV Serosorting

Although a general upward increase in prevalence and incidence has been observed in many US cities, some investigators have noticed a relatively static HIV incidence in certain cities, despite increases in the reported cases of unprotected sexual acts. McConnell and Grant[3] from Gladstone Institute of Virology and Immunology in San Francisco, California, put forward the concept of "HIV serosorting" to partially explain these findings. Essentially, the concept is that HIV-positive people are more likely to have unprotected intercourse with individuals who are also HIV-positive, compared with those who are HIV-negative or of unknown or discordant status. The researchers used a form of sexual networking to establish that this phenomenon existed.

The study presented was a subgroup analysis of an ongoing superinfection study being conducted in San Francisco. Data from 43 HIV-1-infected participants were used in this analysis. Questionnaires that included descriptions of each sexual partner in the last 3 months, serostatus, and types of sexual exposure were completed. Most individuals enrolled had a primary partner. Despite that finding, 65.1% reported at least 1 additional sexual partner in the last 3 months. The mean number of additional partners was 6.3.

The majority (85.7%) of participants reported unprotected anal sexual intercourse with at least 1 other partner. Furthermore, 53.6% had at least 1 partner who was seronegative or of unknown serostatus.

For an analysis of sexual networks, the investigators characterized 176 unique partnerships involving at least 1 HIV-1-infected participant. Some 15.3% of the partnerships involved partners known to be HIV-negative, 36.4% involved partners known to be seropositive, and 48.3% involved partners of unknown serostatus.

In a finding that suggested substantial levels of serosorting, unprotected receptive anal intercourse (URAI) was reported by a partner of a seropositive participant in 7.4% of the serodiscordant partnerships, in contrast to 90.6% of partners in the seroconcordant partnerships. Among partners of unknown serostatus, 34.1% involved URAI. These findings are dramatic and, in some ways, intuitive. The role of serosorting in superinfection remains to be seen.

Of concern remains the high prevalence of unprotected sexual intercourse by HIV-positive persons with partners of unknown serostatus. This highlights the fact that people cannot assume that HIV-positive persons will take universal precautions during sexual activity and that individuals need to assume that all potential partners may be HIV-positive. This paper prompted a lot of questions from the floor, including a statement that HIV incidence was actually increasing in San Francisco. Nonetheless, the occurrence and potential importance of this phenomenon remain clear.

First Evidence of Sexual Transmission From a Long-term Nonprogressor

In the era of highly active antiretroviral therapy (HAART) there is a feeling held among some HIV-positive people that the chance of transmitting HIV sexually will be dramatically reduced in the face of an undetectable viral load. This author concurs with this notion, but with some very important caveats--not the least of which is the potential increase in high-risk sexual behaviors or acquisition of STIs that might offset any potential reduction in transmission induced by HAART. One paper presented at this meeting lent further evidence to the argument that HIV-positive persons are always infectious regardless of viral load.

Mikhail and colleagues[4] presented the first molecular data on the transmission from a long-term nonprogressor with high CD4+ and CD8+ cell counts to 2 individuals, both of whom progressed to AIDS. It has been assumed that this occurs, and it is somewhat surprising that this has not been reported previously.

In this study 15 near full-length molecular clones from the index case and the 2 recipients were generated and compared for sequence diversity. Although analysis of these clones supported the epidemiologic linkage, there were important differences within the nef, tat, and gp41 genes between the donor and recipients. The authors believed that these changes, or "escape mutants," occurred as a result of host-induced selective pressures and probably led to the progression observed in the 2 recipients. Of particular interest was the fact that the viral variants transmitted to the 2 individuals only represented minor variants within the index case, and that these were transmitted despite an undetectable viral load.

In summary, those infected with HIV always remain infectious to others. Furthermore, it cannot be assumed that a virus which runs a benign course within one individual will do so within a different host with a different immune response.

Per Sexual Contact Infection Risk Determined in Serodiscordant Ugandan Couples

If virus can be transmitted by an individual with an undetectable viral load, does it follow that an individual with a high viral load can transmit the virus with even greater efficiency? This is likely and is supported by a number of studies.[5,6]

Wawer and colleagues[7] of Columbia University in New York, NY, added further evidence to this argument by attempting to calculate the per sexual contact risk of contracting HIV with respect to disease stage in the index HIV case.

This group concentrated on serodiscordant couples recruited as part of the Rakai study in Uganda. The index cases were categorized as having acute infections, chronic infections, or advanced AIDS with a very poor clinical condition. A total of 240 monogamous partner pairs were recruited and followed for 10 monthly intervals. The rate of HIV transmission per coital act was stratified by the stage of infection, presence of symptoms, and viral load.

Of the 240 partner pairs, 72 (30%) experienced infection of the other partner over a 4-year period. The rate of HIV transmission per coital act was highest in the 5-month period post seroconversion and was estimated at 0.0081/coital act. This declined to 0.0016/coital act during chronic infection and increased again to 0.0043/coital act in the period before death. Of note, clinical symptoms were not predictive of sexual transmission at any stage. In contrast, the viral load of the index case was the strongest predictor for transmission at every stage of infection.

Data not presented in the abstract were the median viral load of each of the groups studied. However, median viral load of partners with recent infections appeared surprisingly low -- only 13,200 copies/mL. That figure is much lower than other reports of viral loads in persons with well-documented primary HIV infection (before the onset of seroconversion). It begs the question: What would be the per contact transmission risk in true seroconverters? This author believes it would be orders of magnitude higher and would concur with the concept that people during primary HIV infection might cause a disproportionate number of new cases of HIV.

Biology of Transmission

Detection of Primary HIV in Persons With Acute Sexually Transmitted Infections

Pilcher and colleagues[8] from the University of North Carolina, Chapel Hill, presented some fascinating and very disturbing work attempting to calculate the rate of primary HIV infection (PHI) in persons attending STD clinics in Malawi. The research team also tried to determine the predictors of antibody-negative PHI among this cohort. A total of 929 consecutive male patients attending an STD clinic in Malawi were enrolled. All of these patients had acute urethritis and/or genital ulcer disease. A control group of patients was subsequently enrolled from a nearby dermatology clinic (n = 432). All patients were tested for the presence of HIV using rapid latex agglutination tests with positive results confirmed by enzyme-linked immunosorbent assays (ELISA). The resulting negative patients were tested for the presence of HIV RNA using a Roche or NucliSens nucleic acid amplification technique. Positives were confirmed as having PHI by the use of IgM-sensitive ELISA and Western blot. PHI was defined as HIV RNA repeatedly > 10,000 copies/mL with negative/indeterminate antibody test or HIV RNA repeatedly positive with an evolving Western blot. Factors associated with PHI were identified using bivariate analysis and multivariate logistic regression.

Of the 1361 men enrolled, a large number (n = 553; 40.6%) were already HIV-antibody-positive, a reflection of the high HIV prevalence in Malawi. More surprisingly, 24 men (1.8% of the total study population) had confirmed PHI, and 23 of these 24 were STD clinic patients. In contrast to the Rakai study cited above, the median viral load for men with PHI was very high (HIV RNA 6.10 log10 copies/mL [interquartile range (IQR) 5.19-6.54]). This median viral load was significantly higher (P < .0001) than that of the non -PHI HIV-positive men (4.42 log10 copies/mL; IQR 3.91-4.95).

Of note, the factor most strongly associated with PHI diagnosis was presentation at the STD clinic (OR 16.5; 95% confidence interval [CI]: 2.2-122.3). When only the patients from the STD clinic were considered, the 23 PHI cases accounted for 2.5% of all patients, 4.5% of all antibody-negative patients, and 5.0% all HIV infections.

In a multivariate analysis that considered only STD patients, significant predictors of having a PHI diagnosis included diagnosis with first STD (OR 4.96; 95% CI 1.11-22.22); age > 23 years (OR 3.54; 95% CI 1.37-9.13); and education beyond secondary school (OR 2.46; 95% CI 1.03- 5.84). Inguinal adenopathy (OR 4.74; 95% CI 1.97-11.43) was an independent predictor, but genital ulcer disease was not significant in multivariate analysis (OR 2.15; 95% CI 0.86-5.33).

This study is important for several reasons. First, both ulcerative and nonulcerative STDs have been associated with the increased risk of HIV transmission and acquisition. Forty percent of HIV-infected men in this study were coinfected with STDs, and the biological synergy of these concurrent infections might greatly increase the transmission risk that these individuals pose. Even more startling is the number of HIV-antibody-negative persons with primary HIV infection in this cohort. If these individuals were reassured regarding their perceived HIV status, the potential for HIV transmission might be hugely enhanced with the extremely high viral loads documented in these cases. Last, these data suggest that antibody tests alone may be insufficient for excluding HIV infection in STD clinics in the developing world, as their sensitivity was calculated to be less than 95%.

Ultimately, the potential for sexual transmission will depend on not merely the blood plasma viral load, but on the level of viral activity and shedding in the genital secretions as well.

Seminal Super Shedding of HIV-1

In a study by Taylor and colleagues[9] from Birmingham Heartlands Hospital in the United Kingdom, further determinants of seminal HIV shedding were investigated in a cohort of 72 HIV-positive men not taking antiviral therapy.

Clinicians and patients often make erroneous assumptions about the relationship between blood plasma viral load and genital tract viral load. In general, HIV-1 is usually detected in the seminal plasma (SP) less frequently -- and usually at levels approximately 10-fold lower -- than it is in the corresponding blood plasma (BP). However, in most cohorts of patients studied, a significant minority of men have had SP levels that are about equal to, or higher than, corresponding BP levels.

In Taylor and colleagues' study, such individuals were classed as "seminal super shedders" (SSS) of HIV-1, and the study authors postulated that these individuals may represent a group at increased risk of transmitting HIV-1. It is likely that the patients with PHI described above might well fall into the category of SSS. If this is the case, then it might be reasonable to consider that the per sexual contact risk of having sex with one of these individuals may be nearer to 1 than to the 0.008 figures quoted above in Wawer and associates' study.

In this study, HIV-positive men produced matched blood and semen samples at the same time as undergoing tests for sexually transmitted infections. Viral load was determined by Nucleic Acid Sequence-Based Amplification (NASBA). Variables considered were age, CDC status, CD4+ cell count, blood plasma viral load (BLVL), BPVL > 100,000 copies/mL, and the presence or absence of urethritis. SSS were defined as individuals with a seminal plasma viral load (SPVL)/BPVL ratio > 1.

Overall, BPVL was significantly correlated with SPVL (Spearman's sigma 0.53; P < .0001). None of these men had BPVL values < 400 copies/mL. In contrast, 22/72 (31%) had SPVL < 400 copies/mL despite detectable BPVL (ie, they were nonshedders of HIV). Men who shed virus into semen had significantly higher BPVL than did nonshedders (5.01 log10 copies/mL [range, 3.4-6.3] vs 4.2 log10 copies/mL [range, 3.0-5.5]; P < .0001). Nine men met the SSS criteria and these individuals had significantly higher SPVL compared with the main seminal shedding (SS) group (5.6 log10 copies/mL [range, 4.1-6.5] vs 3.9 log10 copies/mL [range, 2.6-6.2]; P < .001). In contrast, BPVL was not significantly different between SSS and SS (4.8 log10 copies/mL vs 5.02 log10 copies/mL, P = .9). The SSS were generally older (48 yrs vs 35 yrs; P < .02), and the presence of urethritis was significantly overrepresented in the SSS group compared with the other groups (3/9 [33%] vs 3/63 [4.8%] of other groups; P = .02). Of note, SSS was not simply explained by high BPVL, as patients with BPVL > 100,000 copies/mL were equally represented in both the SSS group 4/9 (44%) vs SS 25/63 (39%), P > .9. CD4+ cell counts and CDC status were not significantly different between the SSS and SS.

It has been demonstrated previously that nonulcerative STIs can upregulate seminal HIV shedding[10,11]; therefore, it is perhaps unsurprising that urethritis was overrepresented in the SSS. However, it is even more interesting that 6/9 SSS had no evidence of urethritis, and yet they still shed high levels of HIV in their semen. Clearly, further studies into these SSS are required to elucidate possible reasons for the excess shedding, as this important group may pose a higher risk of transmitting HIV to their sexual partners even in the presence of a relatively low blood plasma viral load.

Subclinical Inflammation as a Cause of Increased Vaginal Shedding of HIV

Continuing the theme of genital tract shedding and the effects of genital inflammatory conditions, Lennox and colleagues[12] from Emory University in Atlanta, Georgia, presented data suggesting that subclinical inflammation within the female genital tract might also increase cervicovaginal shedding, as HIV shedding has been described previously with clinically apparent STIs.[13] In their study they defined subclinical inflammation as elevated genital tract levels of IL-1beta or TNF-alpha and attempted to correlate vaginal and plasma viral load with these markers.

The study was conducted from 1996-2000, and blood and vaginal lavages were collected at 987 visits from 135 HIV-infected women enrolled in the Emory Vaginal Ecology Study (EVE). At every/each study visit, each woman was tested for gonorrhea, herpes simplex virus, cytomegalovirus, chlamydia, bacterial vaginosis, and candida vaginitis. Sixty lavage samples were selected and categorized according to the relationship between plasma viral load and genital tract viral load, based on a linear regression analysis of plasma and vaginal viral loads. Twenty samples had genital tract viral load that strongly correlated with plasma viral load; 20 samples had low genital tract viral load compared with plasma viral load (ie, viral load below the lower limit of the 95% confidence interval, around the linear regression line); and 20 samples had high genital tract viral load compared with plasma viral load (ie, viral load above the upper limit of the 95% confidence interval). (This last group of individuals would be categorized as "vaginal super shedders" in the previous study). Each sample was analyzed for IL-1beta and TNF-alpha by commercially available ELISAs.

In these 60 patients, genital tract infections were infrequent and equally distributed among the 3 groups of patient samples. The results shown in the Table give the median plasma viral load (copies/mL) compared with cervicovaginal lavage viral load and the respective concentrations of IL-1beta and TNF-alpha (pg/mL) in the 3 groups of samples.

Table. Correlation Between Vaginal and Plasma Viral Load, and Levels of IL-1beta and TNF-alpha

Sample group Vaginal
viral load
viral load
IL-1beta level TNF-alpha level
< 95% CI
< 100 200,000 7.49 4.00
("normal shedders")
4200 60,000 79 3.14
> 95% CI
470,000 14,000 210 6.98

As shown in the Table, the concentrations of both IL-1beta and TNF-alpha were associated with the vaginal viral loads in the groups of samples (P < .01 and P = .01, respectively, using analysis of variance of ranks).

Of note, the diagnosed genital tract infections did not explain higher or lower vaginal viral levels than would be expected based on plasma viral loads. In contrast, IL-1beta and TNF-alpha levels in vaginal secretions were significantly associated with the vaginal viral loads independent of plasma viral loads. These results suggest that subclinical inflammation, as measured by increased cytokine levels, may be an important determinant of vaginal viral shedding.

Putting these studies together, the clinical relevance becomes clear: High blood plasma viral load means that genital tract viral load is also likely to be high, as the 2 appear to be reasonably correlated. Furthermore, from these and previous published studies, it is likely that clinical or asymptomatic genital inflammation will further increase genital HIV shedding. However, what these studies add to our knowledge is that there appears to be a small number of patients with no apparent genital tract infections who still may have exceptionally high genital tract viral loads and only moderately elevated plasma viral loads. These individuals, together with those individuals experiencing primary HIV infection, are likely to be particularly infectious to others and represent a group of individuals in whom targeted interventions may be particularly efficacious.


  1. Valdiserri RO. Preventing new HIV infections in the U.S.: What can we hope to achieve? Program and abstracts of the 10th Conference on Retroviruses and Opportunistic Infections; February 10-14, 2003; Boston, Massachusetts. Abstract 4.
  2. Chaisson MA, Hirshfield S, Humberstone M, et al. The Internet and high-risk sex among men who have sex with men. Program and abstracts of the 10th Conference on Retroviruses and Opportunistic Infections; February 10-14, 2003; Boston, Massachusetts. Abstract 37.
  3. McConnell J, Grant R. Sorting out serosorting with sexual network methods. Program and abstracts of the 10th Conference on Retroviruses and Opportunistic Infections; February 10-14, 2003; Boston, Massachusetts. Abstract 41.
  4. Mikhail M, Wang B, Beckthold B, Gill JM, Saksena NK. First evidence showing transmission of HIV-1 from a non-progressor to two recipients. Program and abstracts of the 10th Conference on Retroviruses and Opportunistic Infections; February 10-14, 2003; Boston, Massachusetts. Abstract 903.
  5. Pedraza MA, del Romero J, Roldan F, Heterosexual transmission of HIV-1 is associated with high plasma viral load levels and a positive viral isolation in the infected partner. J Acquir Immune Defic Syndr. 1999;21:120-125. Abstract
  6. Colfax GN, Buchbinder SP, Cornelisse PG, Vittinghoff E, Mayer K, Celum C. Sexual risk behaviors and implications for secondary HIV transmission during and after HIV seroconversion. AIDS. 2002;16:1529-1535. Abstract
  7. Wawer MJ, Serwadda D, Li X, et al. HIV-1 transmission per coital act, by stage of HIV infection in the HIV+ index partner, in discordant couples, Rakai, Uganda. Program and abstracts of the 10th Conference on Retroviruses and Opportunistic Infections; February 10-14, 2003; Boston, Massachusetts. Abstract 40.
  8. Pilcher CD, Price MA, Hoffman IF, et al. Frequent detection of acute primary HIV infection in men in Malawi: reconsideration of counseling and testing approaches. Program and abstracts of the 10th Conference on Retroviruses and Opportunistic Infections; February 10-14, 2003; Boston, Massachusetts. Abstract 154.
  9. Taylor S, Sadiq T, Sabin C, et al. Seminal super shedding of HIV: implications for sexual transmission. Program and abstracts of the 10th Conference on Retroviruses and Opportunistic Infections; February 10-14, 2003; Boston, Massachusetts. Abstract 454.
  10. Winter AJ, Taylor S, Workman J, et al. Asymptomatic urethritis and detection of HIV-1 RNA in seminal plasma. Sex Transm Infect. 1999;75:261-263. Abstract
  11. Cohen MS, Hoffman IF, Royce R, et al. Reduction of concentration of HIV-1 in semen after treatment of urethritis: implications for prevention of sexual transmission of HIV-1. AIDSCAP Malawi Research Group. Lancet. 1997;349:1868-1873. Abstract
  12. Lennox J, Ellerbrock T, Bush T, Conley L, Evans-Strickfaden T, Hart C. Subclinical inflammation in the female genital tract is strongly associated with vaginal viral shedding independent of plasma viral load. Program and abstracts of the 10th Conference on Retroviruses and Opportunistic Infections; February 10-14, 2003; Boston, Massachusetts. Abstract 101.
  13. Ghys PD, Fransen K, Diallo MO, et al. The associations between cervicovaginal HIV shedding, sexually transmitted diseases and immunosuppression in female sex workers in Abidjan, Cote d'Ivoire. AIDS. 1997;11:F85-F93. Abstract