Abstract
Background. Secretory leukocyte protease inhibitor (SLPI) is responsible for regulating inflammatory damage to and innate and adaptive immune responses in the vaginal mucosa. Depressed cervicovaginal SLPI levels have been correlated with both Trichomonas vaginalis infection and poor reproductive health outcomes.
Methods. We measured levels of SLPI in 215 vaginal specimens collected from adolescent and young adult females aged 14–22 years. Log-transformed SLPI values were compared by analysis of variance or by an unpaired t test before and after adjustment for confounding effects through the propensity score method.
Results. Females receiving hormonal contraceptives and those with an abnormal vaginal pH had lower SLPI levels as compared to their peers. After propensity score adjustment for race, behavioral factors, hormonal use, and other sexually transmitted infections (STIs), SLPI levels were lower in females with a positive T. vaginalis antigen test result, a vaginal pH >4.5, vaginal leukocytosis, and recurrent (vs initial) T. vaginalis infection, with the lowest levels observed in those with the highest T. vaginalis loads.
Conclusions. The SLPI level was reduced by >50% in a T. vaginalis load–dependent manner. Future research should consider whether identifying and treating females with low levels of T. vaginalis infection (before they become wet mount positive) would prevent the loss of SLPI and impaired vaginal immunity. The SLPI level could be used as a vaginal-health marker to evaluate interventions and vaginal products.
Keywords: SLPI protein, human, Trichomonas vaginalis, Vaginosis, Bacterial, Sexually Transmitted Diseases, Adolescent
Secretory leukocyte protease inhibitor (SLPI) is abundantly produced in human mucosal epithelia, including the female reproductive tract, and plays multiple roles in controlling inflammation and in antibacterial and antiviral innate and adaptive immune responses [1]. An elevated SLPI level appears to protect against human immunodeficiency virus (HIV) acquisition [2], whereas a depressed SLPI level has been associated with failure of vaginal microbicides to protect against HIV acquisition [3, 4].
Depressed cervicovaginal levels of SLPI have been associated with the most common curable vaginal infections: trichomoniasis, caused by sexually transmitted pathogen Trichomonas vaginalis, and bacterial vaginosis [5–7]. T. vaginalis and bacterial vaginosis often coexist [8] and are established risk factors for adverse reproductive outcomes, including pregnancy loss, premature rupture of membranes, and preterm labor [9]. Interestingly, decreased levels of SLPI in amniotic fluid have been associated with premature rupture of membranes [10]. Thus, decreased SLPI levels may be one of the reasons for the poor reproductive outcomes associated with vaginal infections.
Although pregnant women with T. vaginalis infection have an increased risk of poor pregnancy outcomes, antibiotic treatment of T. vaginalis does not alter the risk [11, 12], suggesting that immunologic changes triggered by T. vaginalis infection may persist after parasitological cure. T. vaginalis is known to cause vaginal disruption, as evidenced by an elevated vaginal pH, vaginal leukocytosis, and altered vaginal mucosal immune response [4, 9, 13, 14]. Elevated vaginal pH is associated with sexually transmitted infections (STIs), bacterial vaginosis [15], and poor pregnancy outcomes [16].Vaginal leukocytosis is associated with endometritis [17]. The SLPI level is lower in women infected with T. vaginalis as compared to women with Chlamydia trachomatis infection, Neisseria gonorrhea infection, bacterial vaginosis, or no infection [18]. In vitro, SLPI is digested by T. vaginalis cysteine proteases [19].
At present, there are few studies of SLPI that include large numbers of young nonpregnant females, many of whom use hormonal contraception, which, on its own, has controversial effects on the mucosal immune function in adolescents [20, 21]. In addition, T. vaginalis infections vary from asymptomatic infections detected by nucleic amplification testing only to infections associated with severe vaginal symptoms and a positive result of wet mount testing. Although no lasting acquired immunity has been associated with T. vaginalis [9], initial T. vaginalis infections may behave differently than recurrent or persistent infections in term of innate immune environment. We sought to determine whether vaginal levels of SLPI correlated with T. vaginalis loads, initial versus recurrent T. vaginalis infections, hormonal contraceptive use, and traditional measures of vaginal disruption.
METHODS
Clinical Study Design
This study was part of a larger cross-sectional study assessing the accuracy and acceptability of self-testing for T. vaginalis, the results of which are detailed in our previous work [22]. This analysis and the original study were approved by the hospital's institutional review board. Briefly, adolescent and young adult females ages 14–22 years were recruited from an adolescent clinic and the emergency department. A clinician performed a pelvic examination and collected multiple vaginal swab specimens for immediate testing. One polyester swab was frozen dry at −80°C and stored for future use and testing. Participants completed questionnaires encompassing demographic characteristics, sexual history (including prior STI history, number of sex partners, and use of contraceptives), and vaginal or gynecological symptoms.
Clinical bacterial vaginosis was defined using modified version of the Amsel criteria (ie, an abnormal vaginal pH [>4.5], a high clue cell percentage [>20% of epithelial cells per field], and a positive result of an amine test). Compared with Nugent Gram staining, these objective criteria provide sensitivity and specificity similar to that of the full Amsel criteria [23, 24]. Hormonal contraception in the prior 3 months was characterized as none; combined estrogen and progestin (pills, ring, or patch); or progestin only (depot medroxyprogesterone acetate or progestin-containing intrauterine device).
Laboratory Methods
As described previously [25], vaginal swab specimens were used for wet mount diagnostic testing and for assessment of pH, amines, sialidase, and STIs. Leukocytosis on wet mount testing (magnified at 400×) was categorized as low (≤5 leukocytes per field) or high (≥6 leukocytes per field) [26]. Clue cells (vaginal epithelial cells with adherent bacteria) were categorized as present (>20% of epithelial cells per field) or absent [27]. Vaginal pH (pHydrion, Mikro Essentials Laboratories, Brooklyn, NY) was classified as >4.5 (elevated) or ≤4.5 (normal) [28]. Amines were recorded if a fishy odor was detected after applying potassium hydroxide to the specimen [27]. The wet mount test was deemed positive for T. vaginalis if motile trichomonads were observed on direct microscopy. Sialidase testing (BVBlue, Sekisui Diagnostics, Cambridge, MA), the rapid antigen T. vaginalis test (OSOM T. vaginalis Trichomonas Rapid Test, Sekisui Diagnostics), and T. vaginalis culture (InPouch T. vaginalis culture, BioMed, White City, OR) were performed following manufacturers’ directions. On the basis of prior work, we considered any positive T. vaginalis test to be a true positive T. vaginalis infection [29].
At study closure, the stored vaginal swab sample was brought to room temperature and eluted with 1.0 mL of phosphate-buffered saline for 10 minutes. One half of the eluent was transferred to a vaginal swab collection kit (Aptima, Gen-Probe, San Diego, CA) and transported on ice to the International Sexually Transmitted Diseases Research Laboratory (Baltimore, MD). C. trachomatis and N. gonorrhea were detected using nucleic acid amplification testing (NAAT; Aptima Combo2, Gen-Probe). T. vaginalis and Mycoplasma genitalium were detected on the same NAAT platform, using research-use-only reagents.
The remaining eluent (0.5 mL) was transferred to a sterile vial, stored frozen at −80°C, and shipped frozen to the Laboratory of Genital Tract Biology at Brigham and Women's Hospital, where samples were assayed for total protein content and SLPI. The use of these samples for biomarker analysis was approved by the Brigham and Women's Institutional Review Board for Human Subject Research. The Laboratory of Genital Tract Biology is accredited by the College of American Pathologists for immunoassay analysis and operates under strict quality control procedures. SLPI concentrations were measured by the Quantikine enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN), using a Victor2 reader (Perkin Elmer Life Sciences, Boston, MA), as previously described [30, 31]. For initial screening, all samples were diluted 40-fold in the sample diluent provided by R&D Systems, and all samples showing levels below or above the assay detection range were repeatedly tested at dilutions of 4-fold or 400-fold, respectively, to obtain accurate measurements. All measurements were performed in duplicate. A split quality control pool prepared from cervicovaginal samples was tested on each plate showing an interplate variation of <25%.The total protein concentration in each eluted sample was determined by the BCA assay (Thermo Scientific, Rockford, IL), using the Victor2 counter, and the SLPI concentrations were normalized to milligrams of total protein. For the BCA assay, all samples were screened at a 2-fold dilution in phosphate-buffered saline, and testing was repeated without dilution or at a 5-fold dilution if values were below or above the standard curve, respectively. The coefficient of variation between duplicate values for each sample was <10%.
As the closest surrogate measure of T. vaginalis load available, we used a 4-level grading system based on previously established sensitivities of each T. vaginalis diagnostic test, with the highest load assumed to be corresponding to a positive result wet mount testing (least sensitive test), a moderate load corresponding to a negative result of wet mount testing but positive results of culture and/or antigen testing, and the lowest load corresponding to a positive result of transcription-mediated amplification (TMA) analysis but negative results of all other tests [29, 32, 33]. Theoretically, each of the diagnostic tests can give false-negative results because of testing methods, strain differences, or sample handling. In the literature, the reported false-negative rates are as follows: wet mount testing, 31%–58% [34–36]; culture, 7%–28% [32, 36]; rapid antigen testing, 2%–18% [29, 38]; and TMA analysis, 1%–6% [29, 39, 40]. In addition, the organism load required for a positive test result is reported as 10 000 organism/mL for wet mount testing, 100–500 organisms/mL for culture, 800–2500 organism/mL for rapid antigen testing, and 1–40 organisms/mL for NAAT [41, 42].
Statistical Analyses
Unadjusted Associations
We evaluated the distribution of important covariates by T. vaginalis infection status, using χ2 tests for dichotomous variables and analysis of variance (ANOVA) for categorical variables. SLPI results were not normally distributed, so they were log-transformed (natural log) before comparison by means of the Student t test. Differences in mean log SLPI values are presented as the risk ratio (covariate present vs covariate absent), where a value of >1 represents the degree to which the SLPI level is increased, and a value of <1 indicates the degree of decrease when the covariate is present.
Modeling to Adjust for Confounding
In this observational study, there are many potential confounders, such as age, race, sexual behavior, and types of contraception. Failure to adequately consider these confounding effects may bias the results. For example, African American race is overrepresented in the T. vaginalis–infected group and thus may bias the comparison of SLPI levels between the infected and noninfected females. Traditionally, confounding effects are adjusted for by including confounders in the regression model. However, this approach is limited by modeling assumptions, such as distribution and linear additive assumptions, and often requires increases in sample size as the number of confounders increase. The propensity score method is a relatively new technique that does not require the imposition of any distribution or modeling assumptions and helps boost study power. It was proposed to statistically evaluate causal effects free from confounding effects, by mathematically refashioning an observational study into a randomized study; therefore, studies using the propensity score approach are sometimes considered as pseudo-randomized studies [43]. The most commonly used propensity score techniques are matching and inverse weighting, which is considered more appropriate for small sample sizes. Confounders were predetermined on the basis of the current literature and clinical knowledge. We created propensity scores by using both the matching and the inverse weighting technique. Since they gave the same results, we chose to show the weighted results. The distributions of the confounders were compared before and after application of the propensity score methods. For each hypothesis, propensity scores are first created to balance the distribution of confounders between the exposure of interest (present vs absent groups). Then the groups are compared on the basis of their log SLPI values, adjusted by the inverse weight of propensity scores, using a 2-group t test. To adjust for multiple comparisons, we adopted the false-discovery rate (FDR) procedure [44].
For ease of interpretation and to establish normative ranges in adolescent females, we also present the summary statistics in picograms of SLPI per milligram of total protein, derived by calculating the antilog (exponential) of the log-transformed SLPI levels.
RESULTS
Of the 249 females recruited for the original trichomoniasis study, 215 had samples available for SLPI testing and composed the current study sample. The subjects’ characteristics are summarized in Table 1. The mean age was 18.3 years (range, 14–22 years), and the majority self-identified as African American. T. vaginalis was detected (by any T. vaginalis test) in 53 females (24%), C. trachomatis in 49 (22.6%), N. gonorrhea in 21 (9.7%), and M. genitalium in 30 (14%). Many females had sexual risk factors such as multiple sex partners, a history of prior STI, and a current laboratory-confirmed STI. Females with a positive result of any T. vaginalis test were more likely to have a positive result of an amine test, an abnormal pH, or gonorrhea, compared with those who were negative for T. vaginalis. Of the 53 T. vaginalis infections, 20 (38%) had a high load, 27 (50%) had a moderate load, and 6 (11%) had low load. Additionally, 99 (46%) had no past history of T. vaginalis infection and no current infection and were considered to be T. vaginalis naive; 62 (28.8%) reported prior T. vaginalis infection and were currently uninfected, 21 (9.8%) had current infection only, and 32 (14.9%) had both a past history of and current T. vaginalis infection (these females were considered to have recurrent infection).
Table 1.
Subject Characteristics, by Trichomonas vaginalis (TV) Infection Status
| Variable | Subjects, Proportion (%), by TV Test Result(s) |
|||
|---|---|---|---|---|
| ≥1 Positive | All Negative | Overall | P | |
| African American | 49/53 (92.5) | 138/162 (85.2) | 187/215 (87.0) | .172 |
| Age ≥18 y | 26/53 (49.1) | 88/162 (54.3) | 114/215 (53.0) | .505 |
| History of prior STI | 44/53 (83.0) | 115/162 (71.0) | 159/215 (74.0) | .083 |
| Recent sexa | 13/53 (24.5) | 32/162 (19.8) | 45/215 (20.9) | .458 |
| Vaginal symptoms present | 40/53 (75.5) | 118/162 (72.8) | 158/215 (73.5) | .706 |
| TMA test result | ||||
| C. trachomatis positive | 15/53 (28.3) | 34/162 (21.0) | 49/215 (22.8) | .271 |
| N. gonorrhea positive | 10/53 (18.9) | 11/162 (6.8) | 21/215 (9.8) | .010 |
| M. genitalium positive | 8/53 (15.1) | 22/162 (13.6) | 30/215 (14.0) | .782 |
| High clue cell %b | 15/53 (28.3) | 55/160 (34.4) | 70/213 (32.9) | .415 |
| Any history of douching | 19/53 (35.9) | 51/158 (32.3) | 70/211 (33.2) | .633 |
| Abnormal vaginal pH (>4.5) | 35/52 (67.3) | 81/157 (51.6) | 116/209 (55.5) | .048 |
| Hormonal contraception use | 17/52 (32.7 | 60/155 (37.7) | 77/207 (37.2) | .437 |
| Clinical BV presentc | 7/51 (13.7) | 31/156 (19.9) | 38/207 (18.4) | .325 |
| Multiple sex partners (≥2) | 49/51 (96.1) | 138/149 (92.6) | 187/200 (93.5) | .387 |
| Positive result of amine test | 34/50 (68.0) | 73/158 (46.2) | 107/208 (51.4) | .007 |
| High WBC countd | 12/35 (34.3) | 30/130 (23.1) | 42/165 (25.5) | .177 |
| Sialidase positive (n = 45) | 0/5 (0.0) | 8/40 (20.0) | 8/45 (17.8) | .270 |
Column percentages are calculated on the basis of the total number of subjects with a defined characteristic. Denominators differ because of missing data. Differences were assessed with a χ2 test.
Abbreviations: BV, bacterial vaginosis; C. trachomatis, Chlamydia trachomatis; M. genitalium, Mycoplasma genitalium; N. gonorrhea, Neisseria gonorrhea; STI, sexually transmitted infection; TMA, transcription-mediated amplification; WBC, white blood cell.
a Within the past 2 days.
b Defined as >20% of epithelial cells per field.
c Defined as an abnormal vaginal pH (>4.5), a high clue cell percentage (>20% of epithelial cells per field), and a positive result of an amine test.
d Defined as >6 cells/high-power field on wet mount.
Figure 1 shows the direction and magnitude of each covariate effect on SLPI concentration, unadjusted by propensity score. The bars represent the ratios of mean SLPI concentrations in females with the factor present versus those with the factor absent, illustrating the risk magnitude and direction of change. A risk ratio of <1.0 represents a decrease in SLPI level with the factor present, and a risk ratio of >1.0 represents an increase in SLPI level with the factor present. To interpret the risk ratio, a risk ratio of 0.47 means that SLPI levels are decreased by 53%. This value can also be derived by calculating the antilog (exponential) of the difference between the log-transformed levels. SLPI levels were significantly (P < .05, by t tests) lower for those with any hormonal contraception use, a positive result of a T. vaginalis test (TMA, culture, antigen, or wet mount), an elevated white blood cell (WBC) count on wet mount testing, and an abnormal vaginal pH of >4.5. There was also a trend for SLPI levels to be higher in females >18 years of (ratio 1.36) and lower in females with C. trachomatis infection (ratio 0.64), but the differences did not reach statistical significance (P > .05 and P<.1, respectively).
Figure 1.
Direction and magnitude of covariate effect on the secretory leukocyte protease inhibitor (SLPI) level, unadjusted by propensity score. The bars represent the risk ratios of SLPI concentrations (log10 of pg of SLPI per mg of total protein) in women with the specified factor present vs those with that factor absent. A ratio of <1.0 represents a decrease in SLPI level with the factor present, and a ratio of >1.0 represents an increase in SLPI level with the factor present. Solid dark bars are significant at P < .05, striped bars have borderline significance (P > .05 and ≤.1), and light bars are not significant (P > .1, by the t test). aDefined as >20% of epithelial cells per field. bDefined as an abnormal vaginal pH (>4.5), a high clue cell percentage (>20% of epithelial cells per field), and a positive result of an amine test. cDefined as >6 cells/high-power field on wet mount. Abbreviations: Ag, antigen; BV, bacterial vaginosis; C. trachomatis, Chlamydia trachomatis; STD, sexually transmitted disease; TMA, transcription-mediated amplification; T. vaginalis, Trichomonas vaginalis; WBC, white blood cell.
We compared mean log SLPI concentration across categorical variables (hormonal contraception use, T. vaginalis test results, and T. vaginalis infection history), using ANOVA (Table 2).When hormonal contraception was modeled as a dichotomous variable, any hormonal contraception use was associated with a reduced SLPI concentration; however, as a categorical variable, progestin-only contraception appeared to depress the SLPI level more than a combined estrogen-progestin method (ratio 0.61). The SLPI level was significantly lower for females with a high T. vaginalis load (positive result of wet mount testing, ratio 0.46) as compared to females with negative results of all T. vaginalis tests. Also, recurrent T. vaginalis infection appeared to depress SLPI levels (ratio 0.49) to a greater degree than first-time (current only) T. vaginalis infection.
Table 2.
Effect of Categorical Variables on Mean Secretory Leukocyte Protease Inhibitor (SLPI) Levels
| Variable, Response | Subjects, No. (%) | Mean Log SLPI Levela | Risk Ratio | Pb |
|---|---|---|---|---|
| Hormonal contraception (n = 207) | ||||
| None | 130 (62.8) | 11.287 | Ref | .004 |
| Combined estrogen/ progestin | 31 (15.0) | 11.049 | 0.79 | |
| Progestin only | 46 (22.2) | 10.553 | 0.61 | |
| TV test result(s) (n = 215) | ||||
| All negative | 162 (75.3) | 11.223 | Ref | .006 |
| Only TMA positive | 6 (2.8) | 12.101 | 2.4 | |
| Wet mount negative, antigen or culture positive | 27 (12.6) | 10.720 | 0.60 | |
| Wet mount positive | 20 (9.3) | 10.447 | 0.46 | |
| TV history (n = 215) | ||||
| TV naive | 99 (46.0) | 11.154 | Ref | .011 |
| Prior TV Only | 62 (28.8) | 11.338 | 1.20 | |
| Current TV only | 21 (9.8) | 11.289 | 1.14 | |
| Prior and current TV | 32 (14.9) | 10.435 | 0.49 |
Abbreviations: Ref, reference; TMA, transcription-mediated amplification; TV, Trichomonas vaginalis.
a Data are pg of SLPI per mg of total protein.
b Values are unadjusted and calculated by analysis of variance.
Figure 2 shows the effect of propensity weighting on the distribution of several important covariates, compared with T. vaginalis infection status. Variables displayed are those known to affect either T. vaginalis infection status or SLPI level. The y-axis shows the difference in the percentage of females who were T. vaginalis positive, compared with those who were T. vaginalis negative, before (light bar) and after (dark bar) propensity weighting. For example, before weighting, the proportion of females who self-reported African American race was higher (by about 10%) among those who were T. vaginalis positive as compared to those who were T. vaginalis negative; after weighting, the difference was only about 4%. Although the differences prior to weighting were significant only for N. gonorrhea infection and a positive result of amine testing, all covariates became more evenly distributed after weighting, with no significant differences. In addition to analyzing differences on the basis of T. vaginalis infection status (Figure 2), we also examined the effect of weighting on the distribution of covariates with respect to factors found to be associated with the SLPI level in our data set (ie, abnormal vaginal pH, high WBC count on wet mount testing, and history of T. vaginalis infection; data not shown). In each case, propensity weighting removed the imbalance of variables and resulted in a more even distribution of covariates.
Figure 2.
Effect of propensity weighting on distribution of subject characteristics by Trichomonas vaginalis infection status. The y-axis shows the absolute difference in the percentage of females who were T. vaginalis positive, compared with those who were T. vaginalis negative, within each subject group, before (light bar) and after (dark bar) propensity weighting. For example, before weighting, the percentage of subjects who were African American was higher (by 10%) among those who were T. vaginalis positive, compared with those who were T. vaginalis negative; after weighting, the difference was only about 4%. The weighting completely eliminated the differences that were significant at P < .05 (marked by an asterisk). aDefined as >20% of epithelial cells per field. Abbreviations: CT, Chlamydia trachomatis; GC, Neisseria gonorrhea; MG, Mycoplasma genitalium.
Table 3 shows the results obtained after propensity weighting. After weighting, the SLPI concentration was not associated with the combined variable “positive result of any T. vaginalis test.” The SLPI level was decreased in females with a positive result of the T. vaginalis antigen test (ratio 0.64), an elevated vaginal pH (ratio 0.6), and vaginal leukocytosis (defined as >6 WBCs/HPF on wet mount testing; ratio 0.44). Compared with females with an initial/current infection, females with recurrent T. vaginalis infection had a significantly depressed SLPI level (ratio 0.49). Among females with a positive result of any T. vaginalis test, those with a positive result of wet mount testing had the lowest SLPI level (ratio 0.27), compared with those with a negative result of wet mount testing. After control by use of the FDR procedure, only the T. vaginalis antigen test became marginally nonsignificant (adjusted P = .058 vs unadjusted P = .048); the rest of the results remained the same.
Table 3.
Effect of Variables on Secretory Leukocyte Protease Inhibitor (SLPI) Levels After Propensity Score Weighting
| Variable, Response | Subjects, No. | SLPI Level, Meana |
RRb(95% CI) | Pc | |
|---|---|---|---|---|---|
| Log | Antilog | ||||
| Any TV test result | |||||
| Positive | 53 | 10.92 | 55 271 | 0.79 (.51–1.21) | .27 |
| Negative | 162 | 11.16 | 70 263 | ||
| TV antigen test result | |||||
| Positive | 47 | 10.74 | 46 166 | 0.64 (.41–.995) | .048 |
| Negative | 168 | 11.2 | 73 130 | ||
| Vaginal pH | |||||
| Abnormal (>4.5) | 93 | 10.85 | 51 534 | 0.60 (.40–.89) | .01 |
| Normal (≤4.5) | 116 | 11.36 | 85 819 | ||
| WBC count on wet mount | |||||
| High (>6 cells/HPF) | 42 | 10.44 | 34 201 | 0.44 (.28–.70) | <.01 |
| Normal (≤6 cells/HPF) | 123 | 11.26 | 77 653 | ||
| TV history | |||||
| Naive | 99 | 11.1 | 66 171 | ||
| Past only | 62 | 11.23 | 75 358 | 0.83 (.56–1.37) | .56d |
| Current only | 21 | 11.29 | 80 017 | 0.94 (.51–1.73) | .84e |
| Both past and current | 32 | 10.58 | 39 340 | 0.49 (.30–.91) | .03f |
| Wet mount resultg | |||||
| Positive | 20 | 9.8 | 18 034 | 0.27 (.11–.68) | <.01 |
| Negative | 33 | 11.1 | 66 171 | ||
Analysis is limited to significant variables. Each row represents a separate propensity analysis.
Abbreviations: CI, confidence interval; HPF, high-power field; RR, risk ratio; TV, Trichomonas vaginalis; WBC, white blood cell.
a Log values are propensity score weighted. Antilog values are pg/mg protein and are provided to facilitate interpretation.
b Calculated as [log10 of SLPI level in pg]/[mg total protein] in women with the factor present vs those with the factor absent.
c Two-group t test adjusted by inverse weighting of propensity scores.
d Compared with naive.
e Compared with past only.
f Compared with current only.
g Data are for women with any positive TV test result (n = 53).
DISCUSSION
We have generated solid evidence from these cross-sectional data that a depressed SLPI concentration is highly associated with T. vaginalis infection, as determined by a positive result of rapid antigen testing and wet mount testing. We have previously shown that the rapid antigen test is comparable to culture and more sensitive than wet mount testing [29]. In addition, the antigen test is clinically available, but it has not yet been widely used to assess the risk of outcomes or response to treatment for T. vaginalis.
Our data suggest that the mucosal immune defense is impaired by T. vaginalis in a T. vaginalis load–dependent manner. Among adolescent and young adult females with any evidence of T. vaginalis infection, the SLPI concentration was 73% lower (risk ratio 0.27) in those with positive results of wet mount testing, compared with those with negative results of wet mount testing but positive results of antigen testing, culture, or TMA analysis. However, it is unknown whether identifying and treating women with positive results of NAAT but negative results of wet mount tests or culture (it is presumed that such individuals have low levels of infection) will prevent the decrease in SLPI concentration. This knowledge should be gained through a clinical trial and would be highly useful in determining the appropriateness of asymptomatic T. vaginalis screening, as well as for preventing SLPI-mediated HIV transmission.
It is well known that women with 1 episode of T. vaginalis infection are at high risk for recurrent and persistent infections [45, 46]. A recurrent infection may be facilitated by lower SLPI levels. This hypothesis is in agreement with our finding that the SLPI level was approximately 50% lower among females with recurrent T. vaginalis infection than in those with an initial infection. This is similar to our understanding of the immune response to C. trachomatis infection, in which subsequent infections induce a brisker inflammatory response that damages the host, compared with an initial infection. If these cross-sectional data are confirmed in longitudinal studies, it might lead us to develop a “prevention for positives” approach to the control of trichomoniasis. For example, it may be more cost-effective to develop strategies to prevent future complications among women who test positive for a first episode of trichomoniasis than to prevent the first infection.
Another possible explanation for lower SLPI levels in women with recurrent infection is that some of those women could have a persistent, rather than recurrent, T. vaginalis infection that may not be responding to treatment. A number of recent studies suggest that failed treatment may be more common than was previously reported [45, 47]. We have recently shown that Trichomonas vaginalis can directly suppress SLPI expression via its surface lipophosphoglycan [31].
Finally, we confirmed that traditional markers of vaginal barrier disruption, such as elevated vaginal pH and leukocytosis, are strongly associated with decreases in SLPI concentration and, thus, are markers of impaired vaginal immune defense. This knowledge could have 2 uses. With others, we have shown that women can accurately assess their vaginal pH by using an over-the-counter device and then seek medical care to determine the presence of bacterial vaginosis or T. vaginalis infection, especially if they have engaged in risky behavior [21, 48]. Second, our findings support prior studies suggesting that the SLPI level should be used to evaluate the safety of vaginal products, ranging from tampons to microbicides [49, 50].
Strengths of our study are a reasonable sample size of well-preserved vaginal swab samples from a population of adolescent and young adult females at high risk for T. vaginalis infection and other STIs. The SLPI level was determined using state of the art methods, and levels are standardized by milligrams per total protein concentration, to avoid bias by varying levels of sample on each swab. Our statistical approach, which used propensity weighting, was chosen to minimize the confounding bias inherent in cross-sectional convenience samples. We placed emphasis on the vectorial magnitude of the differences in SLPI levels by presenting risk ratios rather than the absolute values, since at present there are no known cutoffs for normal and abnormal SLPI levels. Our study of 215 predominantly African American adolescents and young adults with a prevalent STI history has limited power to establish a normative range. Therefore, the mean SLPI concentrations presented here only serve as a guide for future studies toward establishing normative ranges for this age category.
The main limitations of our study relate to its cross-sectional design. We relied on self-reporting of past infections, recent sexual contact, and contraception use. Because the SLPI level has a role in alerting the host to a new infection and in modulating both the innate and adaptive immune response, SLPI concentrations may change over time with the duration of infection. With this data set, we could not determine the duration of any T. vaginalis infections. In addition, we could only estimate organism load on the basis of the type of T. vaginalis test yielded positive results. However, our approach is reasonable, given that others have shown by real-time, semiquantitative PCR that culture-positive specimens have a >2 log10 higher concentration of DNA than specimens that are PCR positive but culture negative [42]. Some infections that we labeled as “probable recurrent infections” might have actually represented persistent infections due to failure of prior treatments [45].
In conclusion, in adolescent and young adult females, a depressed SLPI level is strongly associated with T. vaginalis infection, bacterial vaginosis, hormonal contraceptive use, and younger age. Further studies should evaluate the SLPI level as a promising marker of the vaginal immune response to T. vaginalis, bacterial vaginosis, and other STIs in nonpregnant and pregnant females of all ages.
Notes
Financial support. This work was supported by the National Institute of Allergy and Infectious Diseases, National Institutes of Health (NIH; grant 5K23AI063182-04 to J. S. H. and grant 5R01AI079085-04 to R. N. F.), and the National Institute of Biomedical Imaging and Bioengineering, NIH (grant 1U54EB007958-01 to J. S. H.).
Potential conflicts of interest. J. S. H. has received honorarium from Genzyme/Sekisui as a speaker. All other authors report no conflicts of interest.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
- 1.Guerrieri D, Tateosian NL, Maffia PC, et al. Serine leucocyte proteinase inhibitor-treated monocyte inhibits human CD4(+) lymphocyte proliferation. Immunology. 2011;133:434–41. doi: 10.1111/j.1365-2567.2011.03451.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Taborda NA, Catano JC, Delgado JC, Rugeles MT, Montoya CJ. Higher SLPI expression, lower immune activation, and increased frequency of immune cells in a cohort of Colombian HIV-1 controllers. J Acquir Immune Defic Syndr. 2012;60:12–9. doi: 10.1097/QAI.0b013e31824876ca. [DOI] [PubMed] [Google Scholar]
- 3.Ghosh M, Fahey JV, Shen Z, et al. Anti-HIV activity in cervical-vaginal secretions from HIV-positive and -negative women correlate with innate antimicrobial levels and IgG antibodies. PLoS One. 2010;5:e11366. doi: 10.1371/journal.pone.0011366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Fichorova RN, Tucker LD, Anderson DJ. The molecular basis of nonoxynol-9-induced vaginal inflammation and its possible relevance to human immunodeficiency virus type 1 transmission. J Infect Dis. 2001;184:418–28. doi: 10.1086/322047. [DOI] [PubMed] [Google Scholar]
- 5.Pillay K, Coutsoudis A, Agadzi-Naqvi AK, Kuhn L, Coovadia HM, Janoff EN. Secretory leukocyte protease inhibitor in vaginal fluids and perinatal human immunodeficiency virus type 1 transmission. J Infect Dis. 2001;183:653–6. doi: 10.1086/318535. [DOI] [PubMed] [Google Scholar]
- 6.Mitchell CM, Balkus J, Agnew KJ, et al. Bacterial vaginosis, not HIV, is primarily responsible for increased vaginal concentrations of proinflammatory cytokines. AIDS Res Hum Retroviruses. 2008;24:667–71. doi: 10.1089/aid.2007.0268. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Novak RM, Donoval BA, Graham PJ, et al. Cervicovaginal levels of lactoferrin, secretory leukocyte protease inhibitor, and RANTES and the effects of coexisting vaginoses in human immunodeficiency virus (HIV)-seronegative women with a high risk of heterosexual acquisition of HIV infection. Clin Vaccine Immunol. 2007;14:1102–7. doi: 10.1128/CVI.00386-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Brotman RM, Klebanoff MA, Nansel TR, et al. Bacterial vaginosis assessed by gram stain and diminished colonization resistance to incident gonococcal, chlamydial, and trichomonal genital infection. J Infect Dis. 2010;202:1907–15. doi: 10.1086/657320. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Fichorova RN. Impact of T. vaginalis infection on innate immune responses and reproductive outcome. J Reprod Immunol. 2009;83:185–9. doi: 10.1016/j.jri.2009.08.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.King AE, Kelly RW, Sallenave JM, Bocking AD, Challis JR. Innate immune defences in the human uterus during pregnancy. Placenta. 2007;28:1099–106. doi: 10.1016/j.placenta.2007.06.002. [DOI] [PubMed] [Google Scholar]
- 11.Kigozi GG, Brahmbhatt H, Wabwire-Mangen F, et al. Treatment of Trichomonas in pregnancy and adverse outcomes of pregnancy: a subanalysis of a randomized trial in Rakai, Uganda. Am J Obstet Gynecol. 2003;189:1398–400. doi: 10.1067/s0002-9378(03)00777-4. [DOI] [PubMed] [Google Scholar]
- 12.Klebanoff MA, Carey JC, Hauth JC, et al. Failure of metronidazole to prevent preterm delivery among pregnant women with asymptomatic Trichomonas vaginalis infection. N Engl J Med. 2001;345:487–93. doi: 10.1056/NEJMoa003329. [DOI] [PubMed] [Google Scholar]
- 13.Petrin D, Delgaty K, Bhatt R, Garber G. Clinical and microbiological aspects of Trichomonas vaginalis. Clin Microbiol Rev. 1998;11:300–17. doi: 10.1128/cmr.11.2.300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Fichorova RN, Trifonova RT, Gilbert RO, et al. Trichomonas vaginalis lipophosphoglycan triggers a selective upregulation of cytokines by human female reproductive tract epithelial cells. Infect Immun. 2006;74:5773–9. doi: 10.1128/IAI.00631-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Yen S, Shafer MA, Moncada J, Campbell CJ, Flinn SD, Boyer CB. Bacterial vaginosis in sexually experienced and non-sexually experienced young women entering the military. Obstet Gynecol. 2003;102:927–33. doi: 10.1016/s0029-7844(03)00858-5. [DOI] [PubMed] [Google Scholar]
- 16.Simhan HN, Caritis SN, Krohn MA, Hillier SL. Elevated vaginal pH and neutrophils are associated strongly with early spontaneous preterm birth. Am J Obstet Gynecol. 2003;189:1150–4. doi: 10.1067/s0002-9378(03)00582-9. [DOI] [PubMed] [Google Scholar]
- 17.Yudin MH, Hillier SL, Wiesenfeld HC, Krohn MA, Amortegui AA, Sweet RL. Vaginal polymorphonuclear leukocytes and bacterial vaginosis as markers for histologic endometritis among women without symptoms of pelvic inflammatory disease. Am J Obstet Gynecol. 2003;188:318–23. doi: 10.1067/mob.2003.105. [DOI] [PubMed] [Google Scholar]
- 18.Draper DL, Landers DV, Krohn MA, Hillier SL, Wiesenfeld HC, Heine RP. Levels of vaginal secretory leukocyte protease inhibitor are decreased in women with lower reproductive tract infections. Am J Obstet Gynecol. 2000;183:1243–8. doi: 10.1067/mob.2000.107383. [DOI] [PubMed] [Google Scholar]
- 19.Draper D, Donohoe W, Mortimer L, Heine RP. Cysteine proteases of Trichomonas vaginalis degrade secretory leukocyte protease inhibitor. J Infect Dis. 1998;178:815–9. doi: 10.1086/515366. [DOI] [PubMed] [Google Scholar]
- 20.Barousse MM, Theall KP, Van Der Pol B, Fortenberry JD, Orr DP, Fidel PL., Jr Susceptibility of middle adolescent females to sexually transmitted infections: impact of hormone contraception and sexual behaviors on vaginal immunity. Am J Reprod Immunol. 2007;58:159–68. doi: 10.1111/j.1600-0897.2007.00504.x. [DOI] [PubMed] [Google Scholar]
- 21.Scott ME, Ma Y, Farhat S, Shiboski S, Moscicki AB. Covariates of cervical cytokine mRNA expression by real-time PCR in adolescents and young women: effects of Chlamydia trachomatis infection, hormonal contraception, and smoking. J Clin Immunol. 2006;26:222–32. doi: 10.1007/s10875-006-9010-x. [DOI] [PubMed] [Google Scholar]
- 22.Huppert JS, Hesse EA, Bernard MC, Bates JR, Gaydos CA, Kahn JA. Accuracy and trust of self-testing for bacterial vaginosis. J Adolesc Health. 2012;51:400–5. doi: 10.1016/j.jadohealth.2012.01.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Schwebke JR, Hillier SL, Sobel JD, McGregor JA, Sweet RL. Validity of the vaginal gram stain for the diagnosis of bacterial vaginosis. Obstet Gynecol. 1996;88:573–6. doi: 10.1016/0029-7844(96)00233-5. [DOI] [PubMed] [Google Scholar]
- 24.Gutman RE, Peipert JF, Weitzen S, Blume J. Evaluation of clinical methods for diagnosing bacterial vaginosis. Obstet Gynecol. 2005;105:551–6. doi: 10.1097/01.AOG.0000145752.97999.67. [DOI] [PubMed] [Google Scholar]
- 25.Huppert JS, Hesse E, Kim G, et al. Adolescent women can perform a point-of-care test for trichomoniasis as accurately as clinicians. Sex Transm Infect. 2010;86:514–9. doi: 10.1136/sti.2009.042168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Geisler WM, Yu S, Venglarik M, Schwebke JR. Vaginal leucocyte counts in women with bacterial vaginosis: relation to vaginal and cervical infections. Sex Transm Infect. 2004;80:401–5. doi: 10.1136/sti.2003.009134. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Amsel R, Totten PA, Spiegel CA, Chen KC, Eschenbach D, Holmes KK. Nonspecific vaginitis. Diagnostic criteria and microbial and epidemiologic associations. Am J Med. 1983;74:14–22. doi: 10.1016/0002-9343(83)91112-9. [DOI] [PubMed] [Google Scholar]
- 28.Eschenbach DA, Hillier S, Critchlow C, Stevens C, DeRouen T, Holmes KK. Diagnosis and clinical manifestations of bacterial vaginosis. Am J Obstet Gynecol. 1988;158:819–28. doi: 10.1016/0002-9378(88)90078-6. [DOI] [PubMed] [Google Scholar]
- 29.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;45:194–8. doi: 10.1086/518851. [DOI] [PubMed] [Google Scholar]
- 30.Mauck CK, Ballagh SA, Creinin MD, et al. Six-day randomized safety trial of intravaginal lime juice. J Acquir Immune Defic Syndr. 2008;49:243–50. doi: 10.1097/QAI.0b013e318186eae7. [DOI] [PubMed] [Google Scholar]
- 31.Fichorova RN, Lee Y, Yamamoto HS, et al. Endobiont viruses sensed by the human host – beyond conventional antiparasitic therapy. PLOS One. 2012;7:e48418. doi: 10.1371/journal.pone.0048418. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Borchardt KA, Zhang MZ, Shing H, Flink K. A comparison of the sensitivity of the InPouch TV, Diamond's and Trichosel media for detection of Trichomonas vaginalis. Genitourin Med. 1997;73:297–8. doi: 10.1136/sti.73.4.297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.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;49:4106–11. doi: 10.1128/JCM.01291-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Barenfanger J, Drake C, Hanson C. Timing of inoculation of the pouch makes no difference in increased detection of Trichomonas vaginalis by the InPouch TV method. J Clin Microbiol. 2002;40:1387–9. doi: 10.1128/JCM.40.4.1387-1389.2002. 79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Wiese W, Patel SR, Patel SC, Ohl CA, Estrada CA. A meta-analysis of the Papanicolaou smear and wet mount for the diagnosis of vaginal trichomoniasis. Am J Med. 2000;108:301–8. doi: 10.1016/s0002-9343(99)00466-0. [DOI] [PubMed] [Google Scholar]
- 36.Ohlemeyer CL, Hornberger LL, Lynch DA, Swierkosz EM. Diagnosis of Trichomonas vaginalis in adolescent females: InPouch TV culture versus wet-mount microscopy. J Adolesc Health. 1998;22:205–8. doi: 10.1016/s1054-139x(97)00214-0. [DOI] [PubMed] [Google Scholar]
- 37.Wendel KA, Erbelding EJ, Gaydos CA, Rompalo AM. Trichomonas vaginalis polymerase chain reaction compared with standard diagnostic and therapeutic protocols for detection and treatment of vaginal trichomoniasis. Clin Infect Dis. 2002;35:576–80. doi: 10.1086/342060. [DOI] [PubMed] [Google Scholar]
- 38.Huppert JS, Batteiger BE, Braslins P, et al. Use of an immunochromatographic assay for rapid detection of Trichomonas vaginalis in vaginal specimens. J Clin Microbiol. 2005;43:684–7. doi: 10.1128/JCM.43.2.684-687.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Andrea SB, Chapin KC. Comparison of Aptima Trichomonas vaginalis Transcription-Mediated Amplification Assay and BD Affirm VPIII for Detection of T. vaginalis in Symptomatic Women: Performance Parameters and Epidemiological Implications. J Clin Microbiol. 2011;49:866–9. doi: 10.1128/JCM.02367-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Chapin K, Andrea S. APTIMA(R) Trichomonas vaginalis, a transcription-mediated amplification assay for detection of Trichomonas vaginalis in urogenital specimens. Expert Rev Mol Diagn. 2011;11:679–88. doi: 10.1586/erm.11.53. [DOI] [PubMed] [Google Scholar]
- 41.Garber GE. The laboratory diagnosis of Trichomonas vaginalis. Can J Infect Dis Med Microbiol. 2005;16:35–8. doi: 10.1155/2005/373920. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Caliendo AM, Jordan JA, Green AM, Ingersoll J, Diclemente RJ, Wingood GM. Real-time PCR improves detection of Trichomonas vaginalis infection compared with culture using self-collected vaginal swabs. Infect Dis Obstet Gynecol. 2005;13:145–50. doi: 10.1080/10647440500068248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Rubin DB. Estimating causal effects from large data sets using propensity scores. Ann Intern Med. 1997;127:757–63. doi: 10.7326/0003-4819-127-8_part_2-199710151-00064. [DOI] [PubMed] [Google Scholar]
- 44.Benjamini Y, Hochberg Y. On the adaptive control of the false discovery rate in multiple testing with independent statistics. J Educ Behav Stat. 2000;25:60–83. [Google Scholar]
- 45.Peterman TA, Tian LH, Metcalf CA, Malotte CK, Paul SM, Douglas JM., Jr Persistent, undetected Trichomonas vaginalis infections? Clin Infect Dis. 2009;48:259–60. doi: 10.1086/595706. [DOI] [PubMed] [Google Scholar]
- 46.Van Der Pol B, Williams JA, Orr DP, Batteiger BE, Fortenberry JD. Prevalence, incidence, natural history, and response to treatment of Trichomonas vaginalis infection among adolescent women. J Infect Dis. 2005;192:2039–44. doi: 10.1086/498217. [DOI] [PubMed] [Google Scholar]
- 47.Gatski M, Kissinger P. Observation of probable persistent, undetected Trichomonas vaginalis infection among HIV-positive women. Clin Infect Dis. 2010;51:114–5. doi: 10.1086/653443. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Roy S, Caillouette JC, Faden JS, Roy T, Ramos DE. Improving appropriate use of antifungal medications: the role of an over-the-counter vaginal pH self-test device. Infect Dis Obstet Gynecol. 2003;11:209–16. doi: 10.1080/10647440300025523. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Su HI, Schreiber CA, Fay C, et al. Mucosal integrity and inflammatory markers in the female lower genital tract as potential screening tools for vaginal microbicides. Contraception. 2011;84:525–32. doi: 10.1016/j.contraception.2011.02.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Fichorova RN, Yamamoto HS, Delaney ML, Onderdonk AB, Doncel GF. Novel vaginal microflora colonization model providing new insight into microbicide mechanism of action. mBio. 2011;2:e00168–11. doi: 10.1128/mBio.00168-11. [DOI] [PMC free article] [PubMed] [Google Scholar]