TRICHOMONAS VAGINALIS INFECTION AND PROSTATE-SPECIFIC ANTIGEN CONCENTRATION: INSIGHTS INTO PROSTATE INVOLVEMENT AND PROSTATE DISEASE RISK

Abstract

Background:

The protist Trichomonas (T.) vaginalis causes a common, sexually transmitted infection (STI) and has been proposed to contribute to the development of chronic prostate conditions, including benign prostatic hyperplasia and prostate cancer. However, few studies have investigated the extent to which it involves the prostate in the current antimicrobial era. We addressed this question by investigating the relation between T. vaginalis antibody serostatus and serum prostate-specific antigen (PSA) concentration, a marker of prostate infection, inflammation, and/or cell damage, in young, male, U.S. military members.

Methods:

We measured T. vaginalis serum IgG antibodies and serum total PSA concentration in a random sample of 732 young, male U.S. active duty military members. Associations between T. vaginalis serostatus and PSA were investigated by linear regression.

Results:

Of the 732 participants, 341 (46.6%) had a low T. vaginalis seropositive score and 198 (27.0%) had a high score, with the remainder seronegative. No significant differences were observed in the distribution of PSA by T. vaginalis serostatus. However, slightly greater, non-significant differences were observed when men with high T. vaginalis seropositive scores were compared to seronegative men, and when higher PSA concentrations were examined (≥0.70 ng/ml). Specifically, 42.5% of men with high seropositive scores had a PSA concentration ≥0.70 ng/ml compared to 33.2% of seronegative men (adjusted p=0.125).

Conclusions:

Overall, our findings do not provide strong support for prostate involvement during T. vaginalis infection, although our suggestive positive findings for higher PSA concentrations do not rule this possibility out entirely. These suggestive findings may be relevant for prostate condition development because higher early- to mid-life PSA concentrations have been found to predict greater prostate cancer risk later in life.

INTRODUCTION

Trichomonas (T) vaginalis causes the most common, non-viral sexually transmitted infection (STI) and has been proposed to contribute to the development of chronic prostate conditions, such as benign prostatic hyperplasia (BPH) and prostate cancer. Possible mechanisms by which it may contribute to these conditions include: 1) adherence to and lysis of prostate epithelial cells; 2) induction of intra-prostatic inflammation and cytokine production; 3) inhibition of prostate epithelial cell apoptosis; and 4) upregulation of proto-oncogenes.^1,2 More recently, T. vaginalis has also been observed to increase prostate cell line proliferation and invasiveness,^3,4 has been detected in prostate tissue from men with BPH,^2,5 and has been found to be associated with the presence of BPH,⁶ as well as later prostate cancer risk and aggressiveness in some,^7,8 but not all,^9-14 sero-epidemiologic studies.

Many of the original observations of intra-prostatic T. vaginalis infection were made in men infected before the introduction of effective anti-trichomonad therapies (i.e., metronidazole)^15-17 or in men with trichomonal urethritis,¹⁸ a less common manifestation of this frequently asymptomatic infection.¹⁹ Therefore, an important question for prostate condition development is the extent to which trichomonal prostate infection occurs in the current antimicrobial era and in men with asymptomatic infection. We previously addressed this question for chlamydia and gonorrhea by measuring the proportion of young men whose concentration of serum prostate-specific antigen (PSA), a marker of prostate infection, inflammation, and/or cell damage, rose during documented infections.^20,21 However, we were not able to perform a similar type of analysis for T. vaginalis infection because this infection tends to be asymptomatic and is not typically investigated or diagnosed in STI clinics. Therefore, we addressed this question instead by measuring specific serum antibodies to a unique T. vaginalis immunogenic protein as a marker of current and past infection. We investigated the relation between the presence of these antibodies and serum PSA concentration in a study population at high risk for T. vaginalis (i.e., young U.S. military members with a high proportion (45%) of African-American men²²). In addition to informing prostate involvement during infection, our analysis may also inform a possible role for T. vaginalis in prostate carcinogenesis, because higher earlier-life PSA concentrations have been found to be associated with later prostate cancer risk in several studies.^23-26

MATERIALS AND METHODS

Study Population and Design:

Since 1985, the U.S. Armed Forces has conducted routine human immunodeficiency virus (HIV) type-1 testing on all active duty members, as well as pre-service testing and periodic pre- and post-deployment testing. Leftover sera are stored in the Department of Defense Serum Repository (DoDSR) for use in medical research and surveillance.²⁷ With Department of Defense collaboration and approval, sera from the DoDSR can be linked to service members’ electronic military medical record, which includes inpatient and outpatient diagnoses previously coded using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM).

For the present study, we selected a random sample of 750 men with stored specimens in the DoDSR who met the following criteria: were <25 years old as of 1995, were HIV-1 negative, had served on continuous active duty from 1995–2006, and had multiple serum specimens archived within the repository, including one from 2004–6.²⁸ We used these criteria to ensure comparability with our broader parent study of STIs and PSA.^21,29,30 For each of these men, we selected the most recent serum specimen collected from 2004–6. Of the original 750 men selected for this cross-sectional analysis, we excluded 18 with insufficient serum remaining after PSA testing or those whose specimens were left overnight at room temperature during PSA testing, leaving an analytic study sample of 732 men.

Measurement of Trichomonas Vaginalis Serostatus:

We assessed T. vaginalis infection by T. vaginalis serostatus rather than by obtaining diagnoses from the military medical record because T. vaginalis infection is frequently asymptomatic in men and is thus not frequently diagnosed. In addition, even when T. vaginalis infection does present with symptoms, it is not typically diagnosed as T. vaginalis infection, but rather as non-gonococcal urethritis or non-chlamydial, non-gonococcal urethritis (NCNGU). As T. vaginalis infection makes up only 1–20% of non-gonococcal urethritis diagnoses,³¹ we chose to ascertain T. vaginalis exposure by serology rather than by medical record diagnosis to reduce exposure misclassification.

Serum specimens were tested for T. vaginalis antibodies using an in-house enzyme-linked immunosorbent assay that detects IgG antibodies against a recombinant T. vaginalis α-actinin protein.^7,32 Specimens were tested in random order and in duplicate, with inferences determined by the average of the two values (coefficient of variation for the duplicate optical densities (ODs) = 6.6%). A control panel consisting of five quality control specimens of increasing seroreactivity from 0 (no reactivity) to 4+ (strong reactivity) was included in each run (n=6 runs). The average of the control panel values across all runs was then used to determine cut-off points for seropositivity by dividing the average duplicate OD value of the seroreactive control specimens by the average of the 0 reactivity control specimens to create a positive to negative (P/N) ratio. Cut-off points were then derived by taking the mid-point between each P/N ratio – for instance, the cut-off point for the 2+ score was derived by taking the mid-point between the P/N ratios for the 2+ and 1+ control specimens. Scores were assigned to each DoDSR participant by comparing their P/N ratio (i.e., their mean absorbance divided by the average of the 0 reactivity control specimens) to the above-described cut-off points. A score of 0 defined seronegativity, whereas the cut-off points of 1 and ≥2 were used to assign low and high T. vaginalis seropositivity, similar to a contemporary analysis from the same laboratory.⁹

To determine the reproducibility of T. vaginalis serologic testing, we included 19 blinded, duplicate quality control pairs of serum from the DoDSR in the testing sequence (serostatus agreement=73.7% between duplicates).

Measurement of PSA Concentration:

Serum specimens were stored in the DoDSR at –30°C for 0.7–3.7 years and then between –70 and –80°C for 3–15 months until PSA testing.²⁸ PSA remains relatively stable for up to 20 years at temperatures around −20°C.³³ Total PSA concentration was measured using the Beckman Coulter Access Hybritech assay (Beckman Coulter, Brea, CA) at the Johns Hopkins Hospital. PSA was measured rather than abstracted from participants’ military medical records because these men were too young (ages 28 to 36 years at the time of blood draw) for routine PSA screening for prostate cancer based on contemporary guidelines. Specimens were tested in random order and assay reproducibility was determined by testing 25 blinded, quality control pairs from the DoDSR (coefficient of variation=12.4% and 6.9% after excluding one discrepant pair).²⁸

Statistical Analysis:

We compared participants’ demographic, military, and infection history (STI, genitourinary, and any infection) characteristics by T. vaginalis serostatus, using t-tests for continuous variables and chi-square tests for proportions. Infection history was obtained from participants’ military medical record and included the following diagnoses: T. vaginalis infection (ICD-9-CM code 131.02, 131.09, 131.9), NCNGU (099.4, 099.49), genital chlamydia (078.88, 099.3, 099.41, 099.50, 099.53, 099.54, 099.55), gonorrhea (098, 0980, 098.10, 098.11, 098.19, 098.2, 098.89), infectious mononucleosis (075), other STIs such as herpes viruses (054.10, 054.13, 054.19, and any other diagnoses between 090 to 099.9), other genitourinary infections and conditions such as urinary tract infections and BPH (ICD-9-CM codes from 580.0 to 609.0), and any infection or genitourinary condition (STIs [except for T. vaginalis and NCNGU]; infectious mononucleosis; genitourinary infections and conditions; and other non-genitourinary, systemic infections such as infections with cutaneous lesions [022.0, 052.0–053.9, 057.9, 074.0, 085.4], staphylococcus infection [041.1], meningitis and encephalitis [047.9, 062.0], gastro-intestinal tract infections [003, 005.9, 008.4–009.2, 124.0], respiratory tract infections [022.1, 033–038.9, 041.2–041.9, 079.0, 114.0], and other systemic infections [022.9, 078.89, 079.89, 079.99, 084.1, 100.0, 780.6]). All diagnoses from the start of the electronic medical record (1995–7) through to the date of participants’ blood draw (2004–6) were included.

We explored the distribution of PSA concentration by T. vaginalis serostatus by calculating geometric mean, arithmetic mean and median values, and proportions of PSA categories (<0.29, 0.30–0.49, 0.50–0.69, 0.70–0.89, and ≥0.90 ng/ml). We also calculated the percentage of men with PSA values ≥0.70 ng/ml, as this value has been found to be associated with future prostate cancer risk in previous studies of young to middle-aged men.^23,34 Linear regression was used to adjust for potential confounders, including age, race, marital status, military grade, region of residence, reason for blood draw, history and frequency of STIs (excluding T. vaginalis infection and NCNGU), urogenital infections, and systemic infections. As only age, sex, and race influenced the point estimates for T. vaginalis serostatus, only these covariates were retained in the final models. Effect modification by race was investigated by including an interaction term in the regression model and evaluating the significance of this term using the likelihood ratio test.

To determine the possible transient influence of other infections on our findings, we performed sensitivity analyses excluding men diagnosed with any STIs, urogenital, or systemic infections other than T. vaginalis infection and NCNGU within the period of one year before to seven days after specimen collection. We chose this timeframe because we previously observed that large rises in PSA concentration remained up to one year following infection³⁰ and to allow for delayed entry of infectious disease diagnoses into the military medical record (up to 7 days). We also performed sensitivity analyses varying the cut-off points for T. vaginalis seropositivity by using the P/N ratios rather than the mid-points of the P/N ratios as cut-off points. Statistical analyses were performed using SAS version 9.4 (SAS, Cary, NC).

This study was approved by the Institutional Review Boards at the Walter Reed Army Institute of Research and the John Hopkins Bloomberg School of Public Health. All data and specimens were anonymized before release from the DoDSR.

RESULTS

Of the 732 study participants, 341 (46.6%) had a low T. vaginalis seropositive score (score=1), and 198 (27.0%) had a high score (score≥2); the remainder were seronegative. Seropositive men were more likely to be African-American and married than seronegative men, but were similar with respect to all other demographic, military, and clinical characteristics (Table 1).

Table 1:

Demographic, Military, and Infection History Characteristics of Young Male U.S. Military Members by Trichomonas Vaginalis Serostatus, 2004-6

	T. vaginalis Serostatus Score
	0 (n=193)	1 (n=341)	≥2 (n=198)	p-value1	p-value2
Mean Age (years)	33.5	33.5	33.6	0.81	0.70
Race (%)
African-American	40.4	51.6	57.1	0.003	0.001
Caucasian	59.6	48.4	42.9
Marital Status (%)
Married	77.2	84.5	86.4	0.035	0.019
Other	22.8	15.5	13.6
Military Grade (%)
Enlisted	87.6	86.5	86.4	0.93	0.72
Officer	12.4	13.5	13.6
Region of Residence (%)
Northeast	10.4	9.7	11.6	0.89	0.57
Midwest	16.6	15.0	16.2
South	50.8	48.7	44.9
West	9.3	9.7	8.6
Unknown	12.9	17.0	18.7
Reason for Blood Draw (%)
General Force	18.6	19.0	18.7	0.27	0.49
Physical Exam	9.3	14.4	14.1
Part of an STI visit	2.1	0.3	1.5
Other/Unknown	70.0	66.3	65.7
History of (%):
T. vaginalis Infection3	1.0	0.6	1.0	0.67	1.00
T. vaginalis Infection or NCNGU4	3.1	4.1	3.0	0.75	0.96
Genital Chlamydia5	5.2	5.6	6.6	0.83	0.56
Gonorrhea6	1.6	2.6	3.5	0.45	0.34
Infectious Mononucleosis7	1.6	0.6	0.5	0.50	0.37
Urogenital Infections and Conditions8	34.7	27.3	30.8	0.19	0.41
Other STIs9	5.2	3.8	4.0	0.75	0.59
Any STI10	12.4	13.5	13.3	0.94	0.84
Any Urogenital or Systemic Infection11	61.1	58.1	65.2	0.27	0.41

NCNGU=non-chlamydial, non-gonococcal urethritis; STI=sexually transmitted infection.

¹Global p-value

²P-value comparing serostatus 0 and ≥2

³Defined by ICD-9-CM codes: 131.02, 131.09, and 131.9.

⁴Defined by ICD-9-CM codes: 131.02, 131.09, 131.9, 099.4, and 099.49.

⁵Defined by ICD-9-CM codes: 078.88, 099.3, 099.41, 099.50, 099.53, 099.54, and 099.55.

⁶Defined by ICD-9-CM codes: 098, 0980, 098.10, 098.11, 098.19, 098.2, and 098.89.

⁷Defined by ICD-9-CM code: 075.

⁸Defined by ICD-9-CM codes: 580.0 to 609.0. Includes urogenital infections that might contribute to prostate inflammation such as benign prostatitic hyperplasia (600.0), orchitis and epididymitis (604.9), urethritis (597.8), urinary tract infection (599.0), hydrocele (603.9), balanoposthitis (607.1), inflammatory disorders of the penis (607.2), prostatitis (601.0-601.9), and Peyronies disease (607.85).

⁹All STIs (ICD-9-CM codes: 054.10, 054.13, 054.19, and 090 to 099.9), except for genital chlamydia, gonorrhea, T. vaginalis, and NCNGU.

¹⁰All STIs (ICD-9-CM codes: 078.88, 054.10, 054.13, 054.19, and 090 to 099.9), except for T. vaginalis, and NCNGU.

¹¹Includes all STIs (except for T. vaginalis and NCNGU), infectious mononucleosis, urogenital infections and conditions, and non-genitourinary, systemic infections such as infections with cutaneous lesions (022.0, 052.0-053.9, 057.9, 074.0, 085.4); staphylococcus infection (041.1); meningitis and encephalitis (047.9, 062.0); gastro-intestinal tract infections (003, 005.9, 008.4-009.2, 124.0); respiratory tract infections (022.1, 033-038.9, 041.2-041.9, 079.0, 114.0); and other systemic infections (022.9, 078.89, 079.89, 079.99, 084.1, 100.0, 780.6).

No statistically significant differences were observed in the geometric mean, arithmetic mean, median, or categories of PSA concentration between T. vaginalis seropositive and seronegative men in unadjusted or adjusted analyses (Table 2). However, slightly greater, non-significant differences were observed when men with high T. vaginalis seropositive scores were compared to seronegative men, and when PSA concentrations ≥0.70 ng/ml were examined. Specifically, 42.5% of men with high seropositive scores had a PSA concentration ≥0.70 ng/ml compared to 33.2% of seronegative men (adjusted p=0.13). Generally similar results were observed in analyses stratified by race (p_int=0.12), in sensitivity analyses excluding recent infections (41.5% of high seropositive scores with PSA ≥0.70 ng/ml vs. 33.8% seronegative men, adjusted p=0.25), and in sensitivity analyses using the P/N ratios instead of the midpoints as the cut-off points for seropositivity (41.2% of high seropositive scores with PSA ≥0.70 ng/ml vs. 35.2% seronegative men, adjusted p=0.22).

Table 2:

PSA Distribution by Trichomonas Vaginalis Serostatus Among Young U.S. Male Military Members, 2004-6

	T. vaginalis Serostatus Score
	0 (n=193)	1 (n=341)	≥2 (n=198)	p-value1	p-value2
Unadjusted PSA concentration (ng/ml)
Geometric Mean	0.58	0.59	0.62	0.52	0.25
Mean	0.70	0.75	0.73	0.43	0.20
Median	0.55	0.59	0.63	0.38	0.17
Interquartile Range	0.41-0.84	0.39-0.84	0.44-0.88
Full PSA distribution (%)
<0.29	10.4	12.9	9.1	0.43	0.23
0.30-0.49	32.1	25.5	23.2
0.50-0.69	22.8	21.7	23.7
0.70-0.89	14.0	18.5	20.2
≥0.90	20.7	21.4	23.7
≥0.70	34.7	39.9	43.9	0.17	0.06
Adjusted PSA concentration (ng/ml)3
Geometric mean	0.58	0.58	0.61	0.63	0.21
% ≥0.70 ng/ml	33.2	38.8	42.5	0.17	0.13

PSA=Prostate-specific antigen.

¹Global p-value

²P-value comparing serostatus 0 and ≥2

³Adjusted for age, race, and marital status.

DISCUSSION

Overall, no significant associations were observed between T. vaginalis serostatus and PSA concentration in our large cross-sectional sample of young U.S. military members. However, our suggestive, non-significant association between high T. vaginalis serostatus and greater PSA concentrations (≥0.70 ng/ml) leaves open the possibility of an influence of T. vaginalis infection on the prostate.

Comparing our findings to those from studies designed to examine the acute influence of infection on PSA, our non-significant findings are generally consistent with previous null findings for medical record-confirmed NCNGU in the U.S. military,²¹ as well as for those for confirmed gonorrhea and chlamydia when whole shifts in the PSA distribution were examined (i.e., mean and median PSA).^7,21 Our findings differ, however, from previous positive findings for chlamydia and gonorrhea when greater changes in PSA concentrations were examined (≥40% increase above pre-infection values), which we interpreted to mean that only a subset of participants with chlamydia and gonorrhea experienced prostate involvement.^7,21 Although we were not able to examine the distribution of PSA change during T. vaginalis infection in the present study, we did still observe a suggestion of a positive association between T. vaginalis serostatus and higher PSA concentrations, possibly implying prostate involvement during some T. vaginalis infections. Contrary to previous studies, though, this inference is more difficult to make in our analysis because we were not able to measure acute T. vaginalis infection directly but instead had to rely on a marker of cumulative T. vaginalis exposure (i.e., both acute and resolved infections), thereby introducing a certain degree of exposure misclassification. This concern is further exacerbated by possible misclassification of T. vaginalis antibody measurement, although sensitivity analyses using alternative cut-off points yielded generally similar inferences and an expected positive association between T. vaginalis serostatus and African-American race was observed.

Viewing our findings in the context of the longer-term or cumulative influences of current and/or resolved infections rather than the acute influences of infection, our non-significant findings are consistent with those from previous cross-sectional studies of non-lifelong infections (T. vaginalis¹⁴ and syphilis) that observed null associations with PSA,³⁵ but differ from those from other cross-sectional and longitudinal studies that observed positive findings for C. trachomatis serostatus and histories of medical record-confirmed chlamydia and NCNGU diagnoses with PSA concentration and trajectories in younger men.^29,36 However, in the latter study, positive associations were observed only when PSA trajectories were examined and not when only one PSA measurement was used (% with PSA ≥0.7 ng/ml=38.3 for chlamydia, 39.6 for gonorrhea, and 31.7 for NCNGU versus 37.1 for controls without these infections, p-values=0.772, 0.639, and 0.659, respectively²⁹). This latter finding is more in line with our findings and suggests that smaller long-term changes may require longitudinal measurements rather than one cross-sectional measurement to detect as statistically significant. Finally, although our findings differ from previous positive cross-sectional findings for human herpesvirus type 8 serostatus with PSA in older men,^37,38 human herpesvirus type 8 is a chronic, lifelong infection that reactivates periodically across the lifecourse, as opposed to an infection, such as T. vaginalis infection, that can be cleared or cured, which may explain its stronger influence on PSA.

Lastly, if we view our findings in the context of prostate disease development rather than prostate involvement, our non-significant findings are in line with those from a growing body of studies that observed null associations between T. vaginalis serostatus and prostate cancer risk and aggressiveness in older men,^9,11-14 as well as those from one,⁵ but not the other,⁶ previous study of T. vaginalis serology and BPH. However, our findings are not consistent with the high prevalence of T. vaginalis DNA and antigens observed in prostate tissue from older men with BPH.^2,5

CONCLUSIONS

In conclusion, we observed some suggestion of a positive association between T. vaginalis seropositivity and greater PSA concentration. However, this association was not statistically significant. Possible explanations for these non-significant findings include difficulties studying prostate involvement during infection using a cross-sectional study design and detection of antibodies, which does not distinguish between current and past infections, rather than diagnosis of acute infection directly. Future studies may need to address this question by studying changes in PSA during acute infections detected by genitourinary specimen testing in populations at high risk for T. vaginalis and other acute genitourinary infections.