Klinger, Elissa V. MS*; Kapiga, Saidi H. MD, ScD*; Sam, Noel E. MD, MMed†; Aboud, Said MS‡; Chen, Cheng-Yen PhD§; Ballard, Ronald C. PhD§; Larsen, Ulla PhD∧
TRICHOMONAS VAGINALIS IS A COMMON curable sexually transmitted infection (STI) in both women and men. The prevalence of the infection in Africa ranges from 5% to 74% in women and 5% to 29% in men, depending on the populations studied and the diagnostic methods used to detect the infection.1 T. vaginalis is associated with a number of reproductive health morbidities, including vaginitis, cervicitis, urethritis, and pelvic inflammatory disease in women and urethritis and prostatitis in men.1–3 Among pregnant women, T. vaginalis infection has been associated with adverse birth outcomes such as preterm delivery, premature rupture of the membranes, and low birth weight.4–6 In addition to these reproductive health complications, mounting evidence suggests that T. vaginalis may be a cofactor for HIV-1 transmission and acquisition.7–9
Previous studies estimate that approximately 50% to 70% of people infected with T. vaginalis may not manifest symptoms,3,10 a fact that complicates control efforts, especially in resource-scarce settings where laboratory diagnostic capabilities are often deficient. Thus, the majority of people with the disease may not be diagnosed and the prevalence of the infection is unknown in most developing countries. In addition to these challenges, T. vaginalis infection is generally given less emphasis than other STIs and is not a reportable infection. As a result, the infection is infrequently studied and little information exists about the risk factors for it; consequently, its public health importance remains poorly understood. In an effort to better understand the pattern of T. vaginalis infection across a population from sub-Saharan Africa, we conducted an analysis of data from a large community-based study in northern Tanzania to determine the prevalence and risk factors for T. vaginalis infection among women and their regular male partners. The results from this analysis provide information that ultimately can be used in the design of interventions to reduce the burden of T. vaginalis infection among this population.
Study Population, Sampling, and Interviews
We used data from a community-based study designed to determine the risk factors and social implications of infertility and STIs/HIV-1 in Moshi, Tanzania. Women aged 20 to 44 years who were de factor or de jure residents of the sampled households and their regular male partners were eligible to participate in the study. The 2002 population and housing census enumeration areas were used as the primary sampling units. During the first stage of sampling, a total of 150 clusters were selected using probability proportional to the number of women age 20 to 44 to obtain the required number of women aged 20 to 44 years living in the study area. After the clusters were selected, a complete listing of households located in these areas was compiled. In the second stage of sampling, 18 households were randomly selected to participate in the survey from each of the 150 clusters.
Data were collected between November 2002 and March 2003. All eligible women and their partners were interviewed separately using a structured questionnaire after signing a written consent form. The women’s questionnaire collected detailed information about sociodemographic characteristics, pregnancy history, contraceptive use, sexual behavior, STI-related symptoms, infertility, and health care-seeking for infertility problems. The male partner’s questionnaire covered many of the same topics but was less detailed. Every attempt was made to ensure privacy of the respondent by conducting the interviews in a private location within or outside the household with only the respondent and the interviewer present. When the husband/male partner was away from the house, enumerators made an appointment for the male interview to occur at a later time, or they contacted the man at his place of work immediately after the woman’s interview.
After the interview, approximately 5 to 10 mL of blood was collected for HIV-1, herpes simplex virus type 2 (HSV-2), and syphilis testing. Later, respondents were requested to provide a 10- to 15-mL urine specimen in a sterile plastic container for detection of Neisseria gonorrhoeae, T. vaginalis, Chlamydia trachomatis, and Mycoplasma genitalium. Specimens were transported in cold boxes no more than 3 to 4 hours after collection to the Department of Clinical Laboratories at Kilimanjaro Christian Medical Centre (KCMC) for processing and storage. Blood samples were tested at KCMC, whereas urine samples were shipped for testing to the Centers for Disease Control and Prevention (CDC), Atlanta, Georgia. The study protocol was approved by the Ethics Committees of KCMC, the Tanzania National Institute for Medical Research, and the Institutional Review Boards of Harvard School of Public Health, University of Maryland, and the CDC.
Laboratory results, when available, were provided at a health facility or school or other public institution located in the ward where the interview had taken place. Each respondent received results and posttest counseling from the same enumerator who provided pretest counseling and collected the specimens. Confidentiality was maintained by having the enumerators meet with each study subject in a separate room away from other people. Enumerators also provided information about HIV-1 prevention. All respondents who reported symptoms during the interviews or who were diagnosed with laboratory-confirmed STIs received free treatment in accordance with the guidelines of the Tanzania Ministry of Health.
Urine samples were tested for T. vaginalis, C. trachomatis, N. gonorrhoeae, and M. genitalium by using a real-time multiplex polymerase chain reaction (M-PCR) assay. Initially, all urine specimens were pooled in groups of 3 and DNA was extracted using the Qiagen Viral RNA kit (Qiagen), which is recommended for extraction of DNA from urine specimens. DNA was then extracted from all specimens falling within positive pools and tested by the M-PCR assay. Sequence-specific detection of M-PCR amplification products was based on TaqMan technology and was performed using the Rotor-Gene 3,000 instrument (Corbett Research, Australia). M-PCR amplifications were performed in 50-μL reaction tubes using 25 μL of sample DNA. A final concentration of 1× PCR Gold buffer (Applied Biosystems); 4 mmol/L MgCl2; 200 μM each of dATP, dGTP, dCTP, and dUTP were used. Two units of AmpliTaq Gold DNA polymerase (Applied Biosystems) and one unit of uracil-N-glycosylase (UNG; Applied Biosystems) were used per 50-μL reaction.
The amplification target sequences of TaqMan probes and primers were designed from cytosine-specific DNA methyltransferase for N. gonorrhoeae, cryptic plasmid for C. trachomatis, adhesion (MgPa) for M. genitalium, and repeat DNA fragment for T. vaginalis, with amplicon sizes of 80, 87, 79, and 92 bp, respectively. The final concentrations of the forward, reverse primer and the probe for each target were optimized and used between 200 and 400 nM; the performance of M-PCR assay was validated using collections of urine specimens, provided by Dr. David Martin of Louisiana State University in New Orleans and by Dr. Patricia Totten of University of Washington in Seattle, that were previously tested for individual discharge organisms (unpublished data). Amplification was performed using the following parameters: 95°C, 10 minutes and 50°C, 2 minutes, followed by 50 cycles at 95°C for 20 seconds and 60°C for 60 seconds. Positive controls were prepared from cultures using the Qiagen (Valencia, CA) mini-DNA kit for N. gonorrhoeae, C. trachomatis, and T. vaginalis. M. genitalium-positive DNA control was obtained from the ATCC (catalog no. 33530D). The real-time M-PCR assay allows detection over a broad range of target concentrations with analytical sensitivities of one to 10 genomic copies for M. genitalium and N. gonorrhoeae; 0.01 to 0.1 genomic copies for T. vaginalis; and the equivalent of 0.01 to 0.1 inclusion-forming unit for C. trachomatis.
HIV-1 infection was determined by using 2 enzyme-linked immunosorbent assays (ELISAs). Indeterminate or conflicting results were resolved by Western blot (Genetic Systems HIV-1 Western blot; Bio-Rad Laboratories, Redmond, WA). Active or recent syphilis was diagnosed if the serum was reactive to both the rapid plasma reagin card test (Macro-Vue; Becton-Dickinson, Cockysville, MD) and the Treponema pallidum hemagglutination assay (TPHA) (Wellcosyph HA; Murex Biotech Ltd., U.K.). A positive TPHA test alone was interpreted as evidence of a past infection. Antibodies for HSV-2 were detected using type-specific HSV-2 EIA according to the manufacturer’s instructions (HerpeSelect 2 ELISA; Focus Technologies, Cypress, CA).
Separate analyses were performed on the data collected from the sample of women and the sample of men. The frequency distribution of T. vaginalis infection and all potential predictors was inspected. Predictor variables were dummy-coded and their distributions summarized as proportions. The analyses were performed in Stata Version 8.0 (Stata Corp., College Station, TX) taking into account the multistage sample design. The prevalence of T. vaginalis infection in each category of the predictor variables was determined and the associations between T. vaginalis infection and predictor variables were summarized by using odds ratios (ORs) and corresponding 95% confidence intervals (CIs). To adjust for multiple risk factors simultaneously, multivariate analyses were performed by using logistic regression models.11
All variables that were significant in the univariate analyses (P ≤0.05) and others that were thought to be important based on a priori knowledge were included in a list of candidate variables for inclusion in multivariate logistic regression models. Symptoms of STIs were excluded from entry into multivariate logistic regression modeling. The stepwise procedures of forward selection and backward elimination were used to determine the final logistic regression model. After the final models were obtained, we assessed potential confounding of variables not retained in the models by entering one variable at a time into the model. Variables that changed any of the coefficients of the predictors in the final model by >10% were considered potential confounders and were thus retained in the models. The fitness of the final model was checked by using the Hosmer-Lemeshow test.11
Of 2523 households located, 2203 eligible women aged 20 to 44 years were listed in the household questionnaires and 2192 (99.5%) of these women were located and invited to participate in the study. A total of 2019 (92.1%) women who were married or single completed an interview. Among these interviewed women, 563 (27.9%) were not married or had no regular male partner and consequently the husbands or male partners of 1456 (72.1%) women were eligible to participate in the study. Of these eligible male partners, 1233 (84.7%) were located and invited to participate in the study, and a total of 794 (64.4%) were interviewed. Among subjects who participated in the interviews, 1440 married or single women and 588 men who were the regular male partners of the women provided urine specimens for detection of T. vaginalis infection and other STIs. Subsequent analyses included both married and single women but only men who were regular male partners of the women studied.
Male and female subjects who gave a urine specimen were not significantly different from other subjects who did not provide samples in terms of age, duration of stay in Moshi, religion, proportion having children, and the proportion engaging in an income-earning activity. At the time of interview, male and female subjects who provided a urine sample were more likely to report infertility-related problems, alcohol use, and STI-related symptoms and to have a relatively low level of education when compared with subjects who did not provide these samples. Women who provided samples were also more likely to be circumcised and to report high-risk sexual behaviors.
Of 1440 women and 588 men who provided urine specimens, 152 women and 38 men had laboratory-confirmed T. vaginalis infection. The weighted prevalence of T. vaginalis infection among all women was 10.7% (95% CI, 8.5–12.9%) and among men 6.3% (95% CI, 4.0–8.6%). The weighted prevalence of T. vaginalis infection was 12.9% (95% CI, 10.2–15.7%) among single women and 9.3% (95% CI, 7.4–11.3%) among married or cohabiting women. Thus, single women had a higher T. vaginalis infection prevalence than married or cohabiting women, although the difference was nonsignificant. The prevalence of T. vaginalis was not significantly different between married or cohabiting women and their male partners. At the time of interview, less than 20% of women with laboratory-confirmed T. vaginalis reported any symptoms related to this infection, including vaginal discharge, excessive genital secretions, or genital itching. No infected man reported genital discharge or excessive secretions, and only one reported genital itching.
The most prevalent STI was HSV-2 reaching 43.6% (95% CI, 40.4–47.1) for women and 38.9% (95% CI, 34.4–43.7) for men followed by HIV-1 infection rates of 11.2% (95% CI, 8.6–12.0) for women and 6.4% (95% CI, 4.4–9.5) for men. Syphilis (past or present), N. gonorrhoeae, and C. trachomatis infection rates were less than 5.0% for both women and men, and M. genitalium was 3.2% (95% CI, 2.8–4.2) for women and 5.1% (95% CI, 3.3–6.7) for men.
In Table 1, we present the associations between sociodemographic characteristics and T. vaginalis infection. The prevalence of T. vaginalis infection did not vary by age among women, whereas an increasing linear trend with age was observed in men (P = 0.02). The risk of T. vaginalis infection decreased as the educational attainment increased in both women and men (P for linear trend = 0.004 in women and <0.001 in men). Approximately 58% of women and 71% of men in this population reported alcohol consumption. Compared with subjects who never drank alcohol, the risk of T. vaginalis infection was significantly increased among women and men who reported drinking alcohol everyday and among women who drank at least once a week.
Compared with married women, the risk of T. vaginalis infection was significantly increased among women who were separated (OR, 2.92; 95% CI, 1.66–5.13). The majority of women studied (69.1%) had a husband or partner who was older, and only 3.3% of the women in the study were older than their male partners. The risk of T. vaginalis infection generally increased as the age difference between partners widened, with the highest risk among women who were older than their male partners (OR, 4.71; 95% CI, 1.88–11.80). The age difference among women who were older than their partners ranged from 1 to 14 years, with the median difference of 2 years. Other sociodemographic characteristics associated with T. vaginalis infection were having a male partner who had children with other women and among men having no income-earning activity. Women and men who reported problems becoming pregnant were also at increased risk of T. vaginalis infection. Approximately 9% of women and 23.2% of men reported 2 or more sex partners in the last 3 years. Women reporting multiple sex partners had increased risk of T. vaginalis infection, although this was significant only in women who had 2 partners in the last 3 years (OR, 2.19; 95% CI, 1.15–4.19). The majority of women (69.8%) and men (80.4%) had never used condoms during the past year. Use of condoms was not associated with T. vaginalis infection (data not shown).
In Table 2, we present the associations between other STIs and T. vaginalis infection. Women reporting genital itching and who had laboratory-confirmed HSV-2, past syphilis, C. trachomatis, or M. genitalium infections had increased risk of T. vaginalis infection. In men, T. vaginalis was not associated with any of these other STIs. HIV-1 infection was not found to be associated with T. vaginalis infection in either men or women in the univariate analyses.
Information about T. vaginalis infection was available in both spouses of 481 couples interviewed. In these couples, 418 (86.9%) were concordantly T. vaginalis negative, whereas 19 (4.0%) were concordantly T. vaginalis positive. In 44 (9.1%) couples, one member was T. vaginalis-positive, whereas the other member was T. vaginalis-negative (i.e., discordant couples). In both men and women, having a partner infected with T. vaginalis was associated with a 20-fold increased risk of T. vaginalis infection.
In the multivariate analysis, we identified a number of factors independently associated with T. vaginalis infection in this population (Table 3). In both men and women, T. vaginalis infection in the partner was the strongest predictor of infection. Other genital infections independently associated with T. vaginalis among women were HSV-2 and M. genitalium. The risk of T. vaginalis infection remained significantly increased in women with little or no education. Other sociodemographic characteristics associated with increased risk of T. vaginalis infection in women were drinking alcohol everyday, being separated from one’s husband, having problems getting pregnant, and having a husband or partner who had children with other women. Finally, the risk of T. vaginalis infection was significantly reduced among men who had an income-earning activity and for men with at least 9 years of education.
We examined the predictors of T. vaginalis infection among women 20 to 44 years of age and their regular male partners from Moshi Urban District in northern Tanzania. The prevalence of T. vaginalis was 10.7% in women and 6.3% in men, indicating that it was the most prevalent nonviral STI and a major public health problem in this population. Other community-based studies conducted in Tanzania have estimated T. vaginalis prevalence at greater than 20% among women12,13 and 11% in men.14
As observed in other studies,3,10 the majority of T. vaginalis infections in women and all but one infection in men were not associated with self-reported symptoms like genital discharge, itching, or excessive genital secretions. Although these results may be affected by incomplete reporting of symptoms, subclinical presentation of T. vaginalis infection appears to be common in this population. Lack of symptoms associated with the infection poses a significant challenge to prevention efforts within this community, especially given that most health workers in Tanzania use the syndromic approach in the clinical management of STIs.
We found that having a partner infected with T. vaginalis was the strongest predictor of infection. This finding provides support for targeting couples in an effort to reduce the burden of T. vaginalis infection in this population. Such approaches may include increasing couples’ awareness and knowledge about signs and symptoms associated with T. vaginalis and other STIs and prevention strategies like promoting consistent condom use.
Other genital infections were significant predictors of T. vaginalis infection in women. M. genitalium has been associated with nongonococcal urethritis in men15 and cervicitis in women.16,17 The observed associations between M. genitalium and T. vaginalis infection may be the result of both infections sharing the same route of transmission. In addition, the genital tract inflammatory processes associated with M. genitalium infection might facilitate infection with T. vaginalis. The observed associations among C. trachomatis, HSV-2, and T. vaginalis infections may also be the result of the common route of transmission and risk factors. Furthermore, the diminished integrity of the epithelial cells caused by these STIs could facilitate colonization of the mucosa of the urogenital tract by T. vaginalis. These infections were not associated with T. vaginalis in men, possibly because of the smaller sample size of male participants and lack of power. We did not find HIV-1 to be independently predictive of T. vaginalis infection in either women or their male partners.
A number of sociodemographic characteristics were found to be predictive of T. vaginalis infection in our analysis. We showed that low educational attainment was independently predictive of T. vaginalis infection in both men and women. This finding is not likely the result of differences in sexual behavior because the reported number of sex partners and condom use did not vary by the level of education. Subjects with a low level of education may be at increased risk for STIs because of lack of awareness about these infections and limited access to medical services. Men who reported engaging in an income-earning activity had a significantly reduced risk of T. vaginalis infection, possibly reflecting the impact that increased income has on access to education and medical services. It is also possible men engaged in income-earning activity may be at low risk of T. vaginalis as a result of other behavioral factors that were not measures in this study.
Separated women were at an increased risk for T. vaginalis infection, although these women were not more likely than other women in the study to report having multiple sex partners. Similarly, women who self-reported problems with infertility and those who reported that their husbands had children with other women were more likely to be infected with T. vaginalis. The positive relationship between self-reported infertility and T. vaginalis infection is not surprising, because other studies have shown that this infection is associated with increased risk of tubal infertility and pelvic inflammatory disease (PID).18–20 Additionally, underlying coinfection with other STIs that are implicated in PID and that share a common route of transmission with T. vaginalis might account for this observation.21 Having children with other women may be an indirect measure of sexual fidelity of the partners of the interviewed women, although the accuracy of this information is questionable, mainly because many extramarital relationships may not result in having children.
Another independent risk factor of T. vaginalis infection in women was self-reported daily consumption of alcohol, implying that a reduction in alcohol intake can have an impact on the prevalence of T. vaginalis infection in this community. Alcohol consumption can impair judgment and thus may facilitate infection with STIs by lowering the perceived risk of engaging in unprotected sex.22 Alcohol use was not associated with T. vaginalis infection in men.
Prevalence of T. vaginalis infection was higher but not significant in women who reported single status at the time of interview (including divorced, separated, or widowed) in comparison to women who were married or cohabiting with regular male partners. We possibly underestimated the prevalence of T. vaginalis infection among women from our sample as a result of relatively low sensitivity of urine PCR in detecting this infection among women.23,24 Surprisingly, we did not observe any association between self-reported measures of sexual behavior, including number of partners and frequency of condom use, and T. vaginalis infection. This suggests that subjects did not accurately report their sexual histories, thereby introducing potential misclassifications into the study. Given that T. vaginalis is transmitted exclusively through sexual contact, promotion of safer sexual practices, including condom use, should continue to be the main strategies for control of infection within this population.
Our findings should be interpreted in light of the following potential limitations. We analyzed data from 1440 women and 588 men who provided urine specimens. We may have introduced selection bias into our results, because these numbers represent 65.4% and 40.4% of the eligible women and men, respectively. Nevertheless, 2019 women and 794 men agreed to be interviewed, and thus our sample group that provided urine specimens represents 71.3% and 74.1% of all women and men interviewed, respectively. We have no information on STIs or risk factors from those who were eligible but did not participate in the study, and we are unable to assess the potential impact of this bias on our results. Subjects who provided urine samples were more likely to report symptoms consistent with STIs, suggesting that we may overestimate the true prevalence of T. vaginalis infection in this population. Additionally, because men who participated in our study were the regular partners of enrolled women, our findings may not be applicable to men who were not in stable or long-term sexual relationships. Finally, self-reported sexual behaviors might have been affected by social desirability bias.
In summary, our study examines T. vaginalis infection in a large, representative sample of women from Moshi Urban District and their regular male partners. It confirms that this infection is a formidable challenge in this community, where other STIs, including HIV-1, are prevalent. Using laboratory confirmation of STIs and detailed questionnaires, we show that having a partner with T. vaginalis increases risk of infection 20-fold. This finding underscores the importance of treating couples when one partner is found to be infected with T. vaginalis. Furthermore, we have shown that HIV-1 infection is not an independent predictor of T. vaginalis infection, but that other STIs, especially in women, do put one at a substantially higher risk of acquiring T. vaginalis infection.
1.Schwebke JR, Burgess D. Trichomoniasis. Clin Microbiol Rev 2004; 17:794–803.
2.Soper D. Trichomoniasis: Under control or undercontrolled? Am J Obstet Gynecol 2004; 190:281–290.
3.Swygard H, Seña AC, Hobbs MM, et al. Trichomoniasis: Clinical manifestations, diagnosis and management. Sex Transm Infect 2004; 80:91–95.
4.Cotch MF, Pastorek JG II, Nugent RP, et al. Trichomonas vaginalis
associated with low birth weight and preterm delivery: The Vaginal Infections and Prematurity Study Group. Sex Transm Dis 1997; 24:353–360.
5.Klebanoff M, Carey C, Hauth J, et al. Failure of metronidazole to prevent preterm delivery among pregnant women with asymptomatic Trichomonas vaginalis
infection. N Engl J Med 2001; 345:487–493.
6.Minkoff H, Grunebaum AN, Schwartz RH, et al. Risk factors for prematurity and premature rupture of membranes: A prospective study of the vaginal flora in pregnancy. Am J Obstet Gynecol 1984; 150:965–972.
7.Jackson DJ, Rakwar JP, Bwayo JJ, et al. Urethral trichomonas vaginalis infection and HIV-1 transmission. Lancet 1997; 350:1076.
8.Wang C, McClelland S, Reilly M, et al. The effect of treatment of vaginal infections on shedding of HIV-type I. J Infect Dis 2001; 183:1017–1022.
9.Sorvillo F, Kerndt P. Trichomonas vaginalis
and amplification of HIV-1 transmission. Lancet 1998; 351:213–124.
10.Wilkinson D, Abdool Karim SS, Harrison A, et al. Unrecognized sexually transmitted infections in rural South African women: A hidden epidemic. Bull World Health Organ 1999; 77:22–28.
11.Hosmer DW, Lemeshow S. Applied Logistic Regression
. New York: John Wiley & Sons, 1989.
12.Jansen HAFM, Morison L, Mosha F, et al. Geographical variations in the prevalence of HIV and other sexually transmitted infections in rural Tanzania. Int J STD AIDS 2003; 14:274–280.
13.Klouman E, Masenga EJ, Klepp K, et al. HIV and reproductive tract infections in a total village population in rural Kilimanjaro, Tanzania: Women at increased risk. J Acquir Immun Defic Syndr 1997; 14:163–168.
14.Watson-Jones D, Mugeye K, Mayaud P, et al. High prevalence of trichomoniasis in rural men in Mwanza, Tanzania: Results from a population based study. Sex Transm Infect 2000; 76:355–362.
15.Taylor-Robinson D. Mycoplasma genitalium
—An up-date. Int J STD AIDS 2002:145–151.
16.Manhart IF, Dutro SM, Holmes KK, et al. Mycoplasma genitalium
is associated with mucopurulent cervicitis. Int J STD AIDS 2001; 12(suppl 2):69.
17.Uno M, Deguchi T, Komeda H, et al. Mycoplasma genitalium
in the cervices of Japanese women. Sex Transm Dis 1997; 24:284–286.
18.El-Shazly AM, Al-Naggar HM, Soliman M, et al. A study on trichomoniasis and female infertility. J Egypt Soc Parasitol 2001; 31:545–553.
19.Grodstein F, Goldman M, Cramer D. Relation of tubal infertility to history of sexually transmitted diseases. Am J Epidemiol 1993; 137:577–584.
20.Sherman KJ, Daling JR, Weiss NS. Sexually transmitted disease and tubal infertility. Sex Transm Dis 1987; 14:12–16.
21.Paisarntantiwong R, Brockman S, Clarke L, et al. The relationship of vaginal trichomoniasis and pelvic inflammatory disease among women colonized with Chlamydia trachomatis
. Sex Transm Dis 1995; 22:344–347.
22.Mbulaiteye SM, Ruberantwari A, Nakiyingi JS, et al. Alcohol and HIV: A study among sexually active adults in rural southwest Uganda. Int J Epidemiol 2000; 29:911–915.
23.Lawing LF, Hedges SR, Schwebke JR. Detection of trichomonosis in vaginal and urine specimens from women by culture and PCR. J Clin Microbiol 2000; 38:3585–3588.
24.Schwebke JR, Lawing LF. Improved detection by DNA amplification of Trichomonas vaginalis
in males. J Clin Microbiol 2002; 40:3681–3683.