Modeling the Impact of Trichomonas vaginalis Infection on HIV Transmission in HIV-Infected Individuals in Medical Care : Sexually Transmitted Diseases

Secondary Logo

Journal Logo

Original Study

Modeling the Impact of Trichomonas vaginalis Infection on HIV Transmission in HIV-Infected Individuals in Medical Care

Quinlivan, Evelyn Byrd MD*,†,‡; Patel, Shilpa N. MPH*; Grodensky, Catherine A. MPH*; Golin, Carol E. MD‡,¶,‖; Tien, Hsiao-Chuan PhD; Hobbs, Marcia M. PhD*,†,‡,**

Author Information
doi: 10.1097/OLQ.0b013e3182593839
  • Free


The southeastern region of the United States (SE USA) carries an excess burden of the 55,000 human immunodeficiency virus (HIV) diagnoses, and the nearly 19 million sexually transmitted infections (STIs) occurring annually.1,2 In 2007, 46% of new AIDS diagnoses, 40% of the people living with AIDS; and 50% of the 14,000 individuals who died of AIDS lived in the SE.3 Trichomoniasis is the most common curable STI among young sexually active women worldwide, an estimated 174 million new cases of Trichomonas vaginalis (TV) occur globally each year4 and is highly prevalent in African-Americans (AA), a population carrying the burden of HIV in the SE USA.5 In the United States, 3.1% of the adult population is estimated to be infected with TV6; however, detailed epidemiology is limited, as TV is not a reportable infection.

Studies now suggest that trichomoniasis, caused by the protozoan parasite, Trichomonas vaginalis, also plays an important but underappreciated role in increasing the sexual acquisition and transmission of HIV.713 The mechanisms by which TV may increase HIV acquisition include (1) elicitation of an inflammatory response of vaginal, exocervix, and urethral epithelia with recruitment of CD4 lymphocytes, macrophages, and microhemorrhages, potentially compromising the mechanical barrier; (2) association with increased HIV viral load in genital secretions; (3) degradation of secretory leukocyte protease inhibitors; and (4) enhanced susceptibility to bacterial vaginosis or colonization with other abnormal vaginal flora, which in turn increases the risk of HIV acquisition.12

Sexual partners exchange TV readily. One cross-sectional study found 71% of male partners of infected women were infected when assessed at a single time point and several other populations report a high prevalence of TV.14,15 Therefore, it has been hypothesized that TV infection plays a significant role in HIV transmission at a global or national level even if the effect on an individual's risk is small.4

TV remains a highly prevalent STI among HIV-infected patients even when patients are in care for years.16 The effect of TV on HIV transmission from HIV coinfected people is less well defined. HIV acquisition was unchanged in the presence of TV infection in some studies, whereas in others, it was increased by 1.2- to 4.8-fold among cohorts of African and African-American women infected with TV.811 HIV shedding in semen was 19 times higher among 5 men with TV-associated urethritis and 4 times higher in women with TV infection compared with TV-uninfected men and women with HIV.17,18 Three studies have looked at the effect of TV infection on HIV transmission and the effect size ranged from no effect to an increased odds of 1.84.10,19,20 One study used presence or absence of discharge to detect STIs including TV (no effect found). Gray et al reported an effect size of 1.5 but combined TV with other vaginal infections. Only Quinn et al reported the effect (1.8) of TV independently by controlling for other STIs in multivariate analyses.

HIV transmission is a complex event that depends on many interrelated behaviors, including the presence or absence of STIs. Knowledge of population-level HIV transmission can be enhanced less expensively by using individual level data in mathematical modeling. We examined clinical, demographic, and behavioral characteristics associated with TV infection and used this information to conduct mathematical modeling to estimate the number of transmitted HIV infections attributable to TV among a cohort of HIV-infected patients receiving medical care in North Carolina.


Study Site and Participants

The STI substudy was part of 2 HIV prevention studies: (1) Prevention with Positives Special Project of National Significance (SPNS)21 (May 7, 2004 to May 31, 2006) and (2) Centers for Disease Control and Prevention (CDC)-positive STEPS (CDC)22 (May 19, 2004 to October 13, 2004) conducted at the University of North Carolina—Chapel Hill, with Institutional Review Board approval. The UNC Healthcare Infectious Diseases clinic is a public academic medical clinic and was the site of the studies. SPNS eligibility requirements were ≥18 years of age, planning to receive care at the clinic for 1 year, English-speaking, treated at the clinic at least once previously, and currently sexually active or injecting drugs. This study compared provider-based prevention counseling to the combination of specialist delivered motivational interviewing prevention sessions in addition to the provider-delivered prevention standard of care. The CDC study eligibility required the subject to have been diagnosed HIV-positive for 6 months or more before enrollment in addition to the same criteria use for SPNS enrollment. The CDC study evaluated the implementation of provided-delivered prevention counseling in a quasi-experimental pre-post intervention design.23 Subjects who completed the CDC study were allowed to enroll in the SPNS study after a 3-month washout period.

Audio computer-assisted self-interview (ACASI) programs at baseline, 6 and 12 months were used in both studies to record socioeconomic, psychosocial, clinical, and sex behavioral characteristics.21,22 The CDC study collected sexual behavior (number of sex acts, vaginal and anal sex acts, use of condoms, and HIV status of partners) over the prior 3 months, and the SPNS study collected the same sexual behavior over the prior 6 months. In other respects, the data were similar. Chart abstraction was performed for information on zip codes, CD4, HIV RNA, and STI laboratory tests, medical visits, and antiretroviral therapy (ART) use. Interview data collected during participation in the CDC study was used when subjects had participated in both studies. Urine samples were provided at baseline and follow-up (6 months and 13 months). Zip codes were used to approximate rural-urban commuting areas (RUCA), a census-tract–based classification of a residence's rural urban status.24 The following categories were used: urban: RUCA 1 to 3; large town: RUCA 4 to 6; small town: RUCA 7 to 9; rural: RUCA 10.

Conversion of 3-Month Sexual Behavior Data

The only key difference between data collected in the 2 studies was that the referent period for sexual behavior was 3 months for the CDC study and 6 months for the SPNS study. The dually enrolled subjects had both 3-month and 6-month sexual behavior data to use for validation of our method of converting 3- to 6-month data. To estimate CDC participants' 6-month sexual behavior, the number of sex acts in 3 months was multiplied by 2. To estimate the number of sex partners, main partnerships were assumed to remain constant (1 main partner in 3 months is still 1 main partner in 6 months), but casual partners were assumed to increase across time (2 casual partners in 3 months but 4 casual partners in 6 months).25 Participants were considered to have 1 main partner if they were married, in a committed relationship, in a domestic partnership, or lived with a sex partner or spouse. This main partner was subtracted from the total number of sex partners to determine the number of casual partners. Partner numbers for participants with {(n6 months) = 1 + 2(n3 months − 1)} or without {(n6 months) = 2(n3 months)} a main partner were estimated for 6 months. This equation was validated by comparing its results with actual data from 76 participants who enrolled in both studies; the equation was found to slightly underestimate 6-month partner number (estimating 0.91 partners for each actual partner) but did not alter the predicted partner number for any individual.

Laboratory Detection and Treatment of STIs

Nucleic acid amplification tests (NAATs) for the detection of Neisseria gonorrhoeae (NG), Chlamydia trachomatis (CT), and TV were performed on first-catch urine samples collected at baseline, 6 and 12 month follow-up times. For CT/NG testing, we used Gen-Probe APTIMA Combo 2 according to the manufacturer's instructions. For TV testing, we used real-time polymerase chain reaction (PCR) (September 28, 2004 to August 15, 2006)26 or Gen-Probe TV analyte-specific reagents (ASR, after August 15, 2006).27 Published sensitivities and specificities of the TV assays are as follows: 90.1% and 100%, respectively, for detection of TV in urine by real-time PCR.26; 98.2% and 98.0%, respectively, for detection of TV in vaginal swabs by this ASR27; and 95.2% and 98.9%, respectively, for detection of TV in urine by Food and Drug Administration–approved NAAT analogous to the ASR assay.28

Statistical Analyses

Predictors of TV Diagnosis.

Prevalence and incidence estimates were determined from the baseline and the follow-up (6 months and 13 months) STI results, respectively. Participants were considered as TV-infected if TV was detected at any time point. Bivariate analyses determined the associations of hypothesized predictor variables22 with TV diagnosis using χ2 and Wilcoxon rank sum tests. Variables for multivariable logistic regression (performed with and without gender) were selected based on the bivariate analysis and importance in the literature, while avoiding collinear elements.

Estimation of HIV Transmission Attributable to TV.

Analyses to estimate the HIV transmission attributable to the STI study participants and the effect of TV infection on transmission were conducted by estimating the potential number of new HIV infections in 6 months. To create a mathematical model, we adapted an equation used previously21,29 that calculates, based on self-reported risk behavior and other factors, for each study participant, the probability, Pi, that he or she would transmit HIV to a sexual partner. The probability of HIV transmission depends on the number and type of sex acts in 6 months (superscripts a, b, c, d, n, [unprotected]; A, B, C, D, N), [protected], condom effectiveness (β), probability that an unknown partner is HIV infected (λ), HIV viral load (μ), and TV infection (TV). The number of unprotected sex acts in 6 months is represented by lower case superscripts a, b, c, d, and n (a: insertive vaginal, b: receptive vaginal, c: insertive anal, d: receptive anal, and n: total acts). Protected acts are denoted by the corresponding upper case superscript (A, B, C, D, and N). Condom efficacy (β) (when used) was estimated as 90% for the base estimate and 80% to 100% for sensitivity analyses. The probability that a partner of unknown sero-status is actually HIV infected (λ) is also included in the model. The addition of TV-infection effect (TV) and the actual HIV viral load in blood plasma (μ) (instead of a 50% reduction in risk if antiretrovirals were being used) were adaptations to the previously published model.

The effect of TV on HIV transmission is unknown. A meta-analysis of risk estimated the TV effect on HIV acquisition as 1.5 in 2001 and ranged from 0 to 4.8 in subsequent studies of African and AA women.11 The effect on HIV transmission was determined independently in 1 study, reported as 1.8.10 Consequently, the effect of TV on per-act transmission probability was determined by multiplying α by the effects of TV infection (1.8). The TV effect was estimated as 1.0 (no effect) to 2.0 in a sensitivity analysis.

The λ parameter denotes the probability that the partner was already infected with HIV and λ was set as 1 for HIV-positive partners, 0 for HIV-negative partners, 8.47% for HIV unknown status male partners of men30; 0.72% for HIV unknown status male partners of women and 0.22% for HIV unknown status female partners of men.31 The range for the sensitivity analysis for the probability of prior infection was estimated as 50% (1/2) to 200% (2-fold) the predicted prevalence.

The associated per-act transmission probabilities (α) were indicated as αa, αb, αc, and αd (0.0006 for unprotected receptive anal and vaginal intercourse, 0.001 for unprotected insertive vaginal intercourse, and 0.02 for unprotected insertive anal intercourse).21 The terms “receptive and insertive” refer to the role of the study subject in the sex act. The standard per-act transmission probabilities (α) (mentioned later in the text) were estimated from participants who were not on ART and typically have an HIV RNA level of 4.5 log10 copies/mL.10,32

The μ parameter represented the predicted change in transmissibility based on the participants' actual viral load (the majority while on ART) compared with an untreated population average μ/(actual log10 HIV RNA - log104.5) and was estimated as 2.45.9 Most (75%) of the participants were on ART, and viral loads were lower than the base level of 4.5 log10 copies/mL. The fold change effect per log viral load (μ) reduced the impact of the viral load on transmissibility if the subject's viral load was less than the untreated average. In this case, a higher fold-change effect (3.26) will produce a lower estimate of transmission had a greater reduction in the effect of the viral load compared with smaller fold-change effect size (1.86).

The following equation represents the transmission probability. Pi = (1 − λ) {1 − (1−(1 − μ)αaTV)a/n (1 − (1 − μ) αbTV)b/n(1 − (1 − μ)αcTV)c/n (1 − (1 − μ)αdTV)d/n(1−(1 − μ) αaTV)A/n (1 − (1 − μ)αb(1 − β)TV)B/n (1 − (1 − μ)αc (1 − β)TV)C/n(1 − (1 − μ)αd(1 − β)TV)D/n}. Using the probability (Pi) and the number of partners, the number of new HIV infections from each study participant was predicted as the sum of new infections predicted for each of 4 partner types (HIV-negative male, HIV-unknown status male, HIV-negative female, and HIV-unknown status female).



Participants in 2 prevention with positive studies (CDC = 186, SPNS = 234) were offered participation in an STI screening substudy. Of the 186 CDC subjects, 140 agreed to participate in the substudy. Of the 234 SPNS subjects, 223 agreed to participate but 76 of these individuals had previously participated in the CDC substudy, leaving 147 new subjects enrolled from the SPSN study and a total of 287 unique STI substudy subjects. One participant failed to complete the ACASI. Four urine samples contained inhibitors of amplification, leaving a total of 541 usable urine samples from 285 active participants.

The 285 participants included in the STI substudy sample did not differ from those in the total SPNS/CDC study sample in age, sex, or race/ethnicity (all P > 0.05). The mean age was 42 years; 63% were men, 66% were AA 28% were white, 61% had 12 years of education or less, 56% had an annual income of <$10,000, and 61% were unemployed. The study population (37% females, 66% African-American (AA), 28% white non-Hispanic, 6% other race/ethnicity; 31% men who have sex with men (MSM) exposure, 48% heterosexual exposure to HIV, and 63% aged >40 years) was similar to the HIV population in North Carolina (28% women, 66% AA, 27% white non-Hispanic, 7% other, 33% MSM, 25% heterosexual exposure to HIV, and 53% aged >40 years) except that women and older ages were over represented in the substudy. The difference between the 48% heterosexuals in the substudy and 25% heterosexuals in North Carolina (NC) is due to a large category of “no reported risk” in the NC data.

STI Prevalence and Incidence

A total of 541 urine samples were collected from 285 patients and analyzed for NG, CT, and TV (Fig. 1). TV was detected in 34 urine samples from 28 individuals (5 heterosexual men, 0 MSM, and 23 women). TV was detected at baseline in 7.4% (n = 21/285). New TV (incident) infections were found in 3% of patients at the first (6-month) and 2% at the second (13-month) follow-up (5/174 and 2/86). Also TV was present at follow-up secondary to failure to receive treatment (n = 2) and relapse or reinfection (n = 4). CT (3/285 at baseline; 2/174 at first follow-up; 0/86 at second follow-up) and NG (2/285 at baseline; none at either follow-up) were rarely detected, and 3 of these 7 patients were coinfected with TV. All patients with TV, NG, or CT were treated. NG and CT infections were reported the NC Division of Public Health.

Figure 1:
STI screening. Urine was screened using nucleic acid amplification tests to detect Trichomonas vaginalis, Neisseria gonorrhoeae, and Chlamydia trachomatis. BL = baseline; F/u = follow-up.

Bivariate Analyses

Men were less likely to have TV infection compared with women (Table 1; odds ratio [OR], 0.10). Nonwhites were less likely to have TV infection compared with those who were white (OR, 0.28). Participants with higher levels of education or employment were less likely to have TV (OR, 0.24; OR, 0.24, respectively). Participants with HIV RNA copy numbers ≤400 copies/mL were less likely to have TV infection (OR, 0.32). TV infection did vary by type of sexual act and was strongly associated with all vaginal sex acts (protected and unprotected) (Table 2). Privately insured participants were less likely to have TV than uninsured or publicly insured patients (0% vs. 11%). Those living in urban settings had less TV infections than those living in other settings (7% vs. 13%); and those receiving antiretroviral medications versus those who were not (8% vs. 13%), but these differences did not reach statistical significance. The number of sexual partners (>1 vs. <1), alcohol use, and cocaine use were selected because of associations with risks for STI exposure and unprotected sex behaviors but were not associated with TV. TV was not associated with CD4 count, time living with HIV, time in HIV care, overall health status, and distance to clinic (data not shown).

Table 1:
Characteristics of Patients With and Without Trichomonas vaginalis Infection
Table 2:
Trichomonas vaginalis (TV) Infection and Sexual Behaviors

HIV Transmission Risk

TV only contributed to the transmission risk for women, as TV was not detected in MSM and heterosexual men with TV only reported HIV+ partners (Table 3). Using a TV effect size of 1.8, the mathematical model predicted that 0.062 HIV infections will occur per 100 female HIV participants in the absence of TV infection and 0.076 HIV infections will occur if TV infection was prevalent in 22% of the HIV-infected women. HIV infection estimates were 0.062, 0.070, and 0.079 per 100 persons if TV effect sizes 1.0, 1.5, or 2.0 were used, estimating 0%, 14%, or 28% additional HIV infections. In the base estimate, one-fifth of HIV transmission events (22.5%) from HIV-infected women participants will be caused by TV infection if TV remains prevalent in one of 5 women.

Table 3:
Estimated Number of New HIV Infections Transmitted per 100 Persons and Sensitivity Analysis of Factors Impacting Transmission

The model for all subjects was not sensitive to the estimated effect of TV infection or the prevalence of HIV in unknown-status partners reported from the National Health and Nutrition Examination Survey (NHANES) (using ½- to 2-fold).30,31 However, the model was sensitive to the efficacy of condoms (0.14–0.39 per 100 persons using 80%–100% efficacy, ±47% change) on the estimate of HIV infections. The model was also sensitive to the effect size (1.85–3.26) of each log10 change in HIV RNA (−13, +25% of the base estimate). Predicted transmission events ranged from 0.23 per 100 persons with the highest effect size of HIV RNA log change (3.26/(actual HIV RNA − log104.5)) to 0.33 for the lowest effect size for HIV RNA change (1.85/(actual HIV RNA − log104.5)). Most of the HIV RNA values were less than the untreated population average of log104.5, thus a higher effect size produced a larger reduction in transmission risk.


In this study of HIV-infected patients in care, we used individual-level data to inform an attributable risk model to describe the effect TV infection may have on HIV transmissibility. Our model estimates that this range is from 0% to 28%, and the base model estimate is that 22% of projected HIV transmissions from women are attributable to TV infections and up to 2% of all HIV transmission in the United States may be related to TV infection. The estimates of transmission risk in a previous study are approximately 2- to 4-fold higher of that reported here.21 Two differences in the studies may explain this. First, the number of MSM recruited was different (50% vs. 33% in our study). Second, the inclusion of HIV RNA values varied. In the prior study, the transmission risk was estimated as ½ the standard risk if a person was receiving ART, otherwise the transmission risk per act was not adjusted for the viral load. In our study, we adjusted the standard per act transmission risk based on the actual difference between the subject's HIV RNA value and the usual untreated population HIV RNA of 38,000 copies/mL. The model showed that estimates of new HIV infections were sensitive to estimates of transmission probabilities associated with viral load. These results support the recently proposed use of ART as a component of an HIV epidemic control strategy.33,34

The prevalence of TV infection was greater than either gonorrheal or chlamydial infections. This is consistent with other studies of TV prevalence in HIV-infected women,15 where TV infection is reported in 22% HIV + women when culture or PCR are used as the detection methods. This difference is likely to be due to the relative sensitivity of screening for NG and CT compared with traditional methods of screening for TV. Screening for TV has been a manual assay previously and therefore not as readily used for screening due to the labor involved and the low sensitivity of microscopy. The prevalence is only 11% if of wet mounts are used as the detection method. It is also possible that coinfection with HIV allows a more persistent infection with TV but not NG or CT, explaining the increased prevalence. NAAT screening for TV was easily implemented and allowed an unrecognized infection present in >10% of the study population to be identified and treated.

TV infection was associated with high HIV viral loads. HIV RNA was higher in TV-infected participants when examined by either the proportion of participants with <400 copies/mL or by mean HIV copies/mL. The association between high viral loads and TV infection suggests that TV occurs more often in women who are receiving less than adequate care whether they are new to care or only marginally engaged in care. Improved entry and retention in care will prevent complications of both TV and HIV infection and avoid the human and economic costs associated with small, but significant numbers of HIV infections. Other previously described variables associated with TV infection were vaginal intercourse, low socioeconomic status (low educational attainment, unemployment, lack of private insurance), and nonwhite race. Poverty has been consistently associated with TV infection with a 3-fold increase in TV when income was <1.85 of the federal poverty level.6 An association of TV with educational status has been previously reported, but the 5-fold increased risk with low education reported in the current study is much greater.6 The association between low educational attainment and TV suggests that inadequate health literacy may be an important element in the pathway to TV infection.21,22 In data presented here, a 3-fold difference in TV prevalence was seen between whites and nonwhites. However, in multivariable analysis, race remained associated with TV infection but not statistically, perhaps due to sample size or a lack of the racial association within a cohort of HIV-infected people.

The limitations of this study include the self-reported nature of some of the data on which risk estimates are based; however, ACASI technology was used to minimize this. The study was conducted at 1 site, which may limit the applicability to other populations. However, the study site draws patients from a wide geographic area within NC. Demographics of the STI substudy are similar to the demographics of the population of newly HIV-diagnosed persons in North Carolina; therefore, the study findings may be applicable to HIV-infected individuals receiving medical care in North Carolina. Our TV effect estimate of 1.8 is based on HIV acquisition and limited HIV transmission studies. Although these estimates are based on limited data and our analyses would be strengthened by more robust data, there are no other direct data. Moreover, our sensitivity analyses took into account changes in these assumptions. Third, the conversion method used to convert 3-month data to 6-month data did not include an adjustment for changes in sex behavior over time or more than 1 main partner. However, in the subset of patients who were enrolled in both studies, the calculation was able to predict the number of partners subjects reported in 6 months using data from the 3-month data collection. Fourth, models are based on estimates and not observations. A full description of the strengths and weaknesses of HIV transmission risk models are beyond the scope of this article but has been recently discussed.35 Finally, the findings we report here are only based on NAAT testing of urine specimens, other detection methods and use of vaginal specimens instead are likely to influence the findings.

An incidence of TV infection of 2% to 3% for HIV-infected women, as reported here, suggests that >5000 new TV infections are occurring in the 290,000 HIV-infected US women annually and potentially contributing to HIV transmission events. Medical care settings provide the opportunity to address both behavioral and biologic aspects of HIV transmission. Biologic factors, such as TV, that enhance HIV transmission may be addressed with STI and HIV medical treatments. TV diagnostic assays are appropriate when clinical history and examination support use of gonorrhea and chlamydia testing. However, due to the often-asymptomatic nature of TV, the limited use of nucleic acid amplification assays, the insensitivity of cytology, and requirement for a pelvic examination for collection, TV is frequently overlooked. Examination of the synergistic effects of medical care, education, and counseling, combined with STI treatment and antiretroviral use, should provide more insights into effective HIV epidemic control.


1. Aral SO, Fenton KA, Holmes KK. Sexually transmitted diseases in the USA: temporal trends. Sex Transm Infect 2007; 83:257–266.
2. Hall HI, Song R, Rhodes P, et al.. Estimation of HIV incidence in the United States. JAMA 2008; 300:520–529.
3. Centers for Disease Control and Prevention. AIDS in the United States by Geographic Distribution 2009
4. Johnston VJ, Mabey DC. Global epidemiology and control of Trichomonas vaginalis. Curr Opin Infect Dis 2008; 21:56–64.
5. Miller WC, Swygard H, Hobbs MM, et al.. The prevalence of trichomoniasis in young adults in the United States. Sex Transm Dis 2005;32:593–598.
6. Sutton M, Sternberg M, Koumans EH, et al.. The prevalence of Trichomonas vaginalis infection among reproductive-age women in the United States, 2001–2004. Clin Infect Dis 2007; 45:1319–1326.
7. Chesson HW, Blandford JM, Pinkerton SD. Estimates of the annual number and cost of new HIV infections among women attributable to trichomoniasis in the United States. Sex Transm Dis 2004; 31:547–551.
8. Kleinschmidt I, Rees H, Delany S, et al.. Injectable progestin contraceptive use and risk of HIV infection in a South African family planning cohort. Contraception 2007; 75:461–467.
9. McClelland RS, Sangare L, Hassan WM, et al.. Infection with Trichomonas vaginalis increases the risk of HIV-1 acquisition. J Infect Dis 2007; 195:698–702.
10. Quinn TC, Wawer MJ, Sewankambo N, et al.. Viral load and heterosexual transmission of human immunodeficiency virus type 1. N Engl J Med 2000; 342:921–929.
11. Rottingen JA, Cameron DW, Garnett GP. A systematic review of the epidemiologic interactions between classic sexually transmitted diseases and HIV: how much really is known? Sex Transm Dis 2001; 28:579–597.
12. Sorvillo F, Smith L, Kerndt P, et al.. Trichomonas vaginalis, HIV, and African-Americans. Emerg Infect Dis 2001; 7:927–932.
13. van der Pol B, Kwok C, Pierre-Louis B, et al.. Trichomonas vaginalis infection and human immunodeficiency virus acquisition in African women. J Infect Dis 2008; 197:548–554.
14. Sena AC, Miller WC, Hobbs MM, et al.. Trichomonas vaginalis infection in male sexual partners: implications for diagnosis, treatment, and prevention. Clin Infect Dis 2007; 44:13–22.
15. Shafir SC, Sorvillo FJ, Smith L. Current issues and considerations regarding trichomoniasis and human immunodeficiency virus in African-Americans. Clin Microbiol Rev 2009; 22:37–45.
16. Cu-Uvin S, Ko H, Jamieson DJ, et al.. Prevalence, incidence, and persistence or recurrence of trichomoniasis among human immunodeficiency virus (HIV)-positive women and among HIV-negative women at high risk for HIV infection. Clin Infect Dis 2002; 34:1406–1411.
17. Price MA, Zimba D, Hoffman IF, et al.. Addition of treatment for trichomoniasis to syndromic management of urethritis in Malawi: a randomized clinical trial. Sex Transm Dis 2003; 30:516–522.
18. Wang CC, McClelland RS, Reilly M, et al.. The effect of treatment of vaginal infections on shedding of human immunodeficiency virus type 1. J Infect Dis 2001; 183:1017–1022.
19. Deschamps MM, Pape JW, Hafner A, et al.. Heterosexual transmission of HIV in Haiti. Ann Intern Med 1996; 125:324–330.
20. Gray RH, Wawer MJ, Sewankambo NK, et al.. Relative risks and population attributable fraction of incident HIV associated with symptoms of sexually transmitted diseases and treatable symptomatic sexually transmitted diseases in Rakai District, Uganda. Rakai Project Team. AIDS 1999; 13:2113–2123.
21. Morin SF, Myers JJ, Shade SB, et al.. Predicting HIV transmission risk among HIV-infected patients seen in clinical settings. AIDS Behav 2007; 11(suppl):S6–S16.
22. Golin C, Marks G, Wright J, et al.. Psychosocial characteristics and sexual behaviors of people in care for HIV infection: An examination of men who have sex with men, heterosexual men and women. AIDS Behav 2009; 13:1129–1142.
23. Gardner LI, Marks G, O'Daniels CM, et al.. Implementation and evaluation of a clinic-based behavioral intervention: positive steps for patients with HIV. AIDS Patient Care STDS 2008; 22:627–635.
24. US Department of Agriculture. Measuring rurality: rural-urban commuting area codes 2005.
25. Pinkerton SD, Abramson PR, Kalichman SC, et al.. Secondary HIV transmission rates in a mixed-gender sample. Int J STD AIDS 2000; 11:38–44.
26. Hardick J, Yang S, Lin S, et al.. Use of the Roche LightCycler instrument in a real-time PCR for Trichomonas vaginalis in urine samples from females and males. J Clin Microbiol 2003; 41:5619–5622.
27. Huppert JS, Mortensen JE, Reed JL, et al.. Rapid antigen testing compares favorably with transcription-mediated amplification assay for the detection of Trichomonas vaginalis in young women. Clin Infect Dis 2007; 45:194–198.
28. Schwebke JR, Hobbs MM, Taylor SN, et al.. Molecular testing for Trichomonas vaginalis in women: Results from a prospective U.S. clinical trial. J Clin Microbiol 2011; 49:4106–4111.
29. Weinhardt LS, Kelly JA, Brondino MJ, et al.. HIV transmission risk behavior among men and women living with HIV in 4 cities in the United States. J Acquir Immune Defic Syndr 2004; 36:1057–1066.
30. McQuillan GM, Kruszon-Moran D, Kottiri BJ, et al.. Prevalence of HIV in the US household population: the National Health and Nutrition Examination Surveys, 1988 to 2002. J Acquir Immune Defic Syndr 2006; 41:651–656.
31. McQuillan G, Kruszon-Moran D. HIV Infection in the United States household population aged 18–49 years: Results from 1999–2006. NCHS Data Brief. Hyattsville, MD: National Center for Health Statistics, 2008.
32. Wilson DP, Law MG, Grulich AE, et al.. Relation between HIV viral load and infectiousness: A model-based analysis. Lancet 2008; 372:314–320.
33. Quinn T. HIV epidemiology and the effects of antiviral therapy on long-term consequences. AIDS 2008; 22(suppl 3):S7–S12.
34. Granich RM, Gilks CF, Dye C, et al.. Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: A mathematical model. Lancet 2009; 373:48–57.
35. Pinkerton SD. HIV Transmission rate modeling: A primer, review, and extension. AIDS Behav In press.
© Copyright 2012 American Sexually Transmitted Diseases Association