The Prevalence of Trichomonas vaginalis Infection among Reproductive-Age Women in the United States, 2001–2004

Abstract

Trichomonas vaginalis infection is a common sexually transmitted protozoal infection, with an estimated 180 million prevalent cases worldwide [1]. The annual incidence in the United States has been estimated to be 3–5 million cases [2], of which 2 million involve young people (age, 15–24 years) [3]. One recent study estimated that the prevalence of trichomoniasis among the general population of young adult women (age, 18–25 years and above) was 2.8% [4]. However, the prevalence of trichomoniasis among all reproductive-age women is unknown.

T. vaginalis infection is associated with several adverse health outcomes, such as preterm birth or delivery of a low—birth weight infant [5,6–7], symptomatic vaginitis in many women, and facilitation of sexual HIV transmission [8,9,10–11]. As many as 746 new cases of HIV infection among women each year can be attributed to T. vaginalis infection in the United States [12]. The costs associated with trichomoniasis are also significant: the total direct cost attributable to trichomoniasis in the United States was estimated to be US$34.2 million per year in 2000 just among persons aged 15–24 years [13].

The diagnosis of T. vaginalis infection in women is often made by microscopic examination (“wet mount”) of a vaginal fluid specimen or by an incidental finding on a Papanicolaou test report, despite the fact that the sensitivity for these tests may be as low as 50% [14,15,16,17–18]. Other diagnostic options include culture and a DNA probe test, both of which have moderately higher sensitivities [14,15,16,17,18–19]. However, these options are not used routinely and are not cost-effective for many clinicians. PCR is only available in laboratories or research settings, and to date, there is no US Food and Drug Administration—approved PCR test for routine diagnosis of T. vaginalis infection in clinical settings. Diagnosis is further complicated by the fact that as many as 50% of T. vaginalis–infected women are asymptomatic. However, treatment of T. vaginalis infection is simple: administration of 1 dose of metronidazole or tinidazole is effective in >90% of cases [20].

Although T. vaginalis infection is prevalent among sexually active adults and is linked to many adverse outcomes, it is not nationally reportable in the United States. The 2001–2004 National Health and Examination Survey (NHANES) provided the first opportunity to examine the prevalence of trichomoniasis among a nationally representative sample of female subjects aged 14–49 years.

Patients and Methods

Study participants. Participants were reproductive-age women (age, 14–49 years) who participated in NHANES during the period 2001–2004. Informed consent was obtained from all of the NHANES participants aged ⩾18 years or from the parents or guardians of participants aged <18 years. NHANES was approved by the institutional review board of the Centers for Disease Control and Prevention.

Study design and procedures. NHANES uses a complex, stratified, multistage probability sample design with unequal probabilities of selection to obtain a nationally representative sample of the civilian, noninstitutionalized population of the United States [21]. Adolescents, non-Hispanic black persons, and Mexican Americans are oversampled in an effort to obtain a more representative sample and reliable population estimates. All women aged 14–49 years who were selected for NHANES during the period 2001–2004 were potentially eligible for participation in this study of the prevalence of T. vaginalis infection. The demographic data were obtained during the initial home interview process. Race and ethnicity were classified using the following categories: non-Hispanic black, non-Hispanic white, and Mexican American. Women who did not indicate one of these categories were classified as "other" and were included in the population totals. The poverty index ratio was calculated by dividing the total family income by the poverty threshold index, adjusted for family size, for the year of the interview [22]. Data on the level of education attained were limited to subjects aged ⩾18 years and were based on the question of whether high school had been completed. Of the 4646 women aged 14–49 years who were interviewed at home for the 2001–2004 NHANES cycles, 4474 (96%) were seen at the mobile examination center, where they completed an additional private interview using an audio computer-assisted self-interview for the sexual behavior component and a computer-assisted personal interview for the reproductive health component. The full set of NHANES questionnaires, including reproductive health and sexual behavior questions, are available on the NHANES Web site (http://www.cdc.gov/nchs/about/major/nhanes/datalink.htm) [21]. A total of 3754 (84%) of 4474 women who were seen at the mobile examination center consented to perform the self-collected vaginal swab specimen collection.

Laboratory methods. Consenting women were given verbal and written instructions on how to use a cotton-tipped swab to collect their vaginal fluid in the mobile examination center bathroom facility. They were instructed to insert the swab at least 2.5 cm (1 inch) into the vagina, to turn the swab against the walls of the vagina while counting to "ten," and then to place the swab into the collection tube. These swabs were transported to the laboratory facility. Vaginal fluids were extracted from the self-collected vaginal swabs, and these fluids were evaluated for the presence of T. vaginalis by PCR. The sensitivity of the PCR assay exceeds that of wet mount and culture for the detection of T. vaginalis in women; thus, PCR was used as the means of detecting T. vaginalis in this study [14,15,16,17–18]. The methods used have been described elsewhere [23]. In brief, specimens were tested within 1–2 weeks after collection. Positive low-level controls, consisting of 10 trichomonas microorganisms per PCR master mix reaction, were run with each batch of up to 26 specimens. Negative controls were included in each batch. All PCR results positive for T. vaginalis were confirmed by Southern blot hybridization. Results were reported as either positive, negative, or uninterpretable for T. vaginalis.

Statistical analyses. Statistical analyses were performed using SAS software, version 9.1 (SAS Institute), and SAS callable Sudaan (RTI) and account for the NHANES complex survey design by incorporating the survey weights and using a Taylor series linearization to calculate variance estimates. Two sets of NHANES data (data for 2001–2002 and for 2003–2004) were concatenated to establish a 4-year data set, and 4-year examination weights were then created for this analysis [22]. The 4-year examination weights were used to account for the unequal probabilities of selection and adjustment for nonresponse and to provide unbiased estimates for the sample of women who were tested. Because there were some missing laboratory specimens, we investigated whether any additional nonresponse adjustments to the original NHANES weights were needed. Certain variables, such as higher education level, non-Hispanic white race/ethnicity, and ever having sex, were associated with an increased likelihood that a laboratory specimen would be provided (P < .05; data not shown). To further evaluate this, we examined the weighted NHANES estimates with an additional nonresponse and poststratification adjustment and found that prevalence estimates remained within the 95% CI based on the original NHANES weights. Therefore, no additional adjustments have been made for nonresponse. Population estimates were generated by multiplying population counts obtained from the postcensal estimates of the civilian, noninstitutionalized population as of 1 July 2002 by the weighted prevalence estimate [24].

Ninety-five percent CIs were calculated using a log-transformation [25]. Significance tests for associations between the presence of T. vaginalis and other categorical variables were based on a Wald χ² statistic; a .05 level defined statistical significance. To test a linear trend across the categories of an independent variable, logistic regression was used, treating the categorical variable as a continuous variable.

Logistic regression modeling using Sudaan was used to study the independent association between the prevalence of trichomoniasis and multiple demographic and behavioral variables among all women tested. Continuous variables, such as age at the time of examination, were treated as continuous if the assumption of linearity with the log odds of trichomoniasis seemed to be reasonable, and variables with >2 categories were collapsed when the prevalence estimates were similar and the context was logical. Variables that excluded subgroups of women on the basis of the skip pattern of the questionnaire were not included in the model, such as questions that were asked only of sexually active women. Goodness of fit for the logistic regression model was assessed using the Hosmer-Lemeshow goodness-of-fit Satterthwaite adjusted F test.

We identified all variables that had a P value ⩽.10 in bivariate analysis. Using a backwards elimination approach, variables were removed in the order of least significance at each step, until all of the remaining variables were significant (P ⩽ .05). Statistical significance in the logistic regression model was based on the P value from the Satterthwaite adjusted F test. When all variables in the model were statistically significant, all pairwise interactions in the model were explored. A pairwise interaction was retained only if the overall P value for the interaction was ⩽.05.

Results

Overall, 3754 (84%) of 4474 participating women who were seen at the mobile examination center provided swab specimens that were evaluable. The overall prevalence of T. vaginalis was 3.1% (95% CI, 2.3%–4.3%); among non-Hispanic white women, the prevalence was 1.3% (95% CI, 0.7%–2.3%); among Mexican American women, it was 1.8% (95% CI, 0.9%–3.7%); and among non-Hispanic black women, it was 13.3% (95% CI, 10.0%–17.7%) (table 1). Other sociodemographic characteristics (in addition to race/ethnicity) that were significantly associated with presence of T. vaginalis included having a high school education or less, poverty, and being born in the United States (table 1). The increasing prevalence with age was particularly notable among non-Hispanic black women (figure 1), among whom the prevalence increased from 8.3% among 14–19-year-old subjects to almost 20% among 40–49-year-old women (P = .001). Estimates involving age for the other racial/ethnic groups were unreliable (relative standard error, 50%); thus, all other racial/ethnic groups were combined for a more stable estimate, and no increase with age was observed.

Table 1

Prevalence of Trichomonas vaginalis infection, by selected sociodemographic characteristics, among girls and women in the United States aged 14–49 years, 2001–2004.

Open in new tabDownload slide

Figure 1

Prevalence of Trichomonas vaginalis among non-Hispanic black women, compared with the prevalence among women from all other racial/ethnic groups, by age, 2001–2004. *P < .001, by Wald adjusted F trend test for age among non-Hispanic black subjects; P = .57 for all other racial/ethnic groups.

Open in new tabDownload slide

The prevalence of T. vaginalis infection was also associated with sexual, reproductive, and personal hygiene characteristics (table 2). Increasing numbers of lifetime sex partners (P < .001, by χ² test for trend), increasing numbers of recent sex partners (P = .001, by χ² test for trend), and early initiation of sexual activity (at age 9–15 years) were associated with a higher prevalence of T. vaginalis infection.

Table 2

Prevalence of Trichomonas vaginalis infection, according to selected sexual, reproductive health, and behavioral characteristics, among girls and women aged 14–49 years in the United States, 2001–2004.

Open in new tabDownload slide

The prevalence of T. vaginalis infection did not differ between women who reported the following symptoms and those who did not: vaginal discharge (4.8% vs. 3.1%; P = .3), vaginal itching (2.6% vs. 3.3%; P = .6), or vaginal odor (6.1% vs. 3.1%; P = .3). Among all women with T. vaginalis infection, 9.5% reported experiencing discharge, 7.3% reported experiencing itching, 7.0% reported having odor, and 15.2% reported having at least 1 vaginal symptom. Of women who had data available for both T. vaginalis infection and bacterial vaginosis, 49.8% (95% CI, 39.9%–62.3%; P = .002) of those who had T. vaginalis infection also had bacterial vaginosis. There were no significant differences in reported symptoms among women with and without T. vaginalis infection (data not shown).

In multivariate analysis, independent risk markers and risk factors for diagnosis of trichomoniasis are shown in table 3 and include race/ethnicity, older age at the time of screening (this was a continuous variable showing increasing risk for each year of age), birth in the United States, fewer years of education, douching in the past 6 months, and a higher lifetime number of sex partners.

Table 3

Multivariate analysis of factors associated with prevalent Trichomonas vaginalis infection (n = 3274).

Open in new tabDownload slide

Discussion

In this nationally representative sample, the prevalence of trichomoniasis among 14–49-year-old women in the United States was 3.1%, corresponding to 2.3 million women with trichomoniasis (95% CI, 1.7–3.2 million women with trichomoniasis). Within the same years of the NHANES national survey data sets, the prevalences of 2 other sexually transmitted diseases, Neisseria gonorrhea and Chlamydia trachomatis infection, were 0.33% and 2.5%, respectively, among female persons aged 14–39 years [26]. Our data show that T. vaginalis infection is a highly prevalent sexually transmitted infection among reproductive-age girls and women aged 14–49 years. In addition, our data provide new information on a wider age group of nationally representative women, compared with the previously reported estimated prevalence of trichomoniasis of 2.8% among 18–25-year-old and above women [4].

Our study revealed that a disproportionately high proportion of non-Hispanic black women had T. vaginalis infection (13.3%), compared with non-Hispanic white women (1.3%) and Mexican American women (1.8%). These differences persisted but were reduced after we controlled for many sociodemographic, sexual, and behavioral characteristics. This finding may indicate that the sex partners of non-Hispanic black women also had a higher prevalence of T. vaginalis infection than did the sex partners of women in other racial/ethnic groups [27]. Although studies for selected clinic populations have also found that the prevalence of trichomoniasis among non-Hispanic black women was higher than that among women from other racial/ethnic groups [28,29–30], this study is the first, to our knowledge, to describe this finding in a nationally representative sample of reproductive-aged women. The reasons for such a large disparity are unclear. The racial/ethnic differences in the prevalence of T. vaginalis infection and the fact that T. vaginalis infection increases the likelihood of sexual transmission of HIV may, in part, explain the disproportionate rate of HIV infection among non-Hispanic black women. Additional research to explore improved trichomoniasis prevention efforts, as well as increased use of point-of-care screening and improved treatment efforts in places where the rate of coinfection is high, could have an impact on the prevalences of both HIV infection and T. vaginalis infection.

We also found that the prevalence of T. vaginalis infection increases with age among non-Hispanic black women, even after controlling for other variables (figure 1). This was also found in a study conducted among women who attended an urban health sexually transmitted diseases clinic [31]. This pattern differs from the epidemiology of other acute sexually transmitted infections, such as chlamydia and gonorrhea; for those infections, the prevalence is highest among adolescents and young adults [32, 33]. Our findings raise several questions. Are the T. vaginalis infections detected among women aged >30 years prevalent or incident infections? Is current treatment for T. vaginalis infection effective at eradicating infection, and are there age-associated differences in treatment efficacy? In the absence of treatment, how long can T. vaginalis persist in the female genital tract? From whom do older women acquire T. vaginalis infection? Are older men more likely to have trichomoniasis than younger men? Because 85% of the infections we detected were asymptomatic, the true duration of the infection is unknown. The number of recent sex partners was also a risk factor for infection with T. vaginalis, although estimates stratified by age and number of recent sex partners were unstable. Practitioners should be aware that, unlike other sexually transmitted infections, the prevalence of T. vaginalis infection does not appear to decrease with age, and many women do not have symptoms. Univariate analyses revealed that certain feminine hygiene practices, such as douching and using feminine powder, were significantly associated with T. vaginalis infection. These hygiene practices have negative effects on vaginal microflora and, in turn, may increase the risk of acquiring trichomoniasis. Previous pregnancy was also noted to be significantly associated with T. vaginalis infection; additional studies are needed to better understand this association.

A strength of this study was the use of PCR to detect T. vaginalis. Nucleic acid amplification tests are the most sensitive tests for detecting organisms and perform well to detect T. vaginalis in different laboratories [14,15,16,17–18]. Published PCR sensitivities range from 81%–97% [14,15,16,17,18–19], exceeding the sensitivities of wet mount, Papanicolaou test, DNA probe, or culture. However, one limitation of PCR is that it is not possible to distinguish between viable and nonviable organisms. Our finding that symptoms were not significantly associated with trichomonas makes it clear that this is often an asymptomatic infection and suggests that screening, even among asymptomatic women, may be an important mechanism of diagnosing this infection that is tied to several important adverse reproductive health outcomes.

Our study also has some limitations. First, we lacked information on recent treatment for trichomoniasis. Such information might have provided some insight into previously treated infection that could have resulted in a positive PCR result without viable organisms. Second, we did not ascertain the performance of the trichomonas PCR used in this study, because we did not use a comparative, parallel diagnostic reference test for trichomoniasis, such as culture. The performance of PCR for detection of T. vaginalis infection has been evaluated and validated in several studies [14,15,16,17–18].

Targeting prevention efforts to decrease the ongoing sexual transmission of T. vaginalis infection may require increased screening efforts. Because most women were asymptomatic, enhanced screening efforts could lead to increased awareness about this infection and to treatment and reduction of transmission. Newly available rapid screening tests for T. vaginalis, including those that involve self-collected vaginal swabs for use outside of clinical settings, are accurate and might improve screening efforts [34,35,36,37–38]. In addition, integration of rapid testing into routine sexually transmitted disease evaluation and care may be a way of accessing persons who may be at risk and need screening quickly and inexpensively [39].

In summary, we found a prevalence of T. vaginalis infection of 3.1% in the United States, and infection disproportionately affected non-Hispanic black women—a group that is also disproportionately impacted by HIV infection, as shown in other studies [40]. Optimal prevention and control strategies for T. vaginalis infection should be further explored as a means of closing the racial disparity in prevalence and decreasing other adverse health outcomes associated with this sexually transmitted infection.

Acknowledgments

We thank Dr. Jeanne Jordan (University of Pittsburgh and Microbiology and Molecular Diagnostics Laboratory, Magee Women's Hospital) for laboratory testing for T. vaginalis. We also thank the NHANES study teams and participants.

Financial support. Division of STD Prevention, National Center for HIV, STD, and TB Prevention, Centers for Disease Control and Prevention (CDC); the National Center for Health Statistics, CDC; and the US Public Health Service, Department of Health and Human Services.

Potential conflicts of interest. All authors: no conflicts.

References