Prospective Study of Trichomonas vaginalis Infection and Prostate Cancer Incidence and Mortality: Physicians' Health Study

Jennifer R. Stark; Gregory Judson; John F. Alderete; Vasanthakrishna Mundodi; Ashwini S. Kucknoor; Edward L. Giovannucci; Elizabeth A. Platz; Siobhan Sutcliffe; Katja Fall; Tobias Kurth; Jing Ma; Meir J. Stampfer; Lorelei A. Mucci


J Natl Cancer Inst. 2009;101(20):1406-11. 

In This Article


In this large nested case–control study, we provide further evidence to support the previously reported association between a T vaginalis–seropositive status and prostate cancer risk.[12] The magnitude of the overall association of T vaginalis–seropositive status with incidence in our study, although not statistically significant, was similar to that observed in the previous case–control study nested in the Health Professionals Follow-up Study (OR = 1.43, 95% CI = 1.00 to 2.03). The Health Professionals Follow-up Study found a suggestion that infection was primarily associated with more aggressive disease, as shown by the higher Gleason scores at diagnosis, but small numbers prohibited a subgroup analysis among men with advanced disease. In this analysis with more than two decades of follow-up for case subjects with prostate cancer, we found that T vaginalis–seropositive status was primarily associated with clinically relevant prostate cancer. That is, compared with a seronegative status, a seropositive status before cancer diagnosis was associated with a statistically significant risk of developing prostate cancer that was diagnosed at an advanced stage. Moreover, T vaginalis infection appears to be associated with cancer that will ultimately progress to bony metastases and prostate cancer death, independent of body mass index, smoking status, aspirin randomization group, age at diagnosis, and tumor stage and grade. We found no evidence of a stronger association with higher Gleason grade but the subjectivity of Gleason grading and the shift in scores over time[16–18] could explain this discrepancy, because Gleason scores in the Health Professionals Follow-up Study tended to be assigned more recently and, thus, may be better predictors of lethal disease.[18]

Our study had several limitations. Because all men provided blood samples in 1982 and all T vaginalis assays of plasma samples were completed in 2008, the performance of the assay should not be differentially influenced by specimen quality according to date of cancer diagnosis. The unknown period of time between infection and blood collection, however, could influence assay sensitivity. Presumably, men who were infected with T vaginalis closer to the time of blood collection in 1982 would be more likely to have detectable levels of antibodies. Because case and control subjects were matched on age (range = 40–84 years at blood collection) and timing of infection is more likely to be related to age than calendar time, this misclassification would likely be nondifferential with respect to case–control status and thus lead us to underestimate the true effect estimate.

Two additional biases also warrant attention. First, we found that the association between T vaginalis infection and incidence of prostate cancer was stronger among men diagnosed within 5 years of blood collection. Biomarkers most strongly associated with disease occurring early in a study typically raise concerns about reverse causation (ie, because of the influence of early preclinical disease on the measured biomarker). However, in this study and in all studies of prostate cancer, biological heterogeneity and the impact of PSA testing on the type of prostate cancers diagnosed are important considerations. Consequently, the men who were diagnosed with prostate cancer earlier in our follow-up, before the introduction of PSA testing in 1986, are more likely to be clinically relevant. Thus, the association observed among case subjects who were diagnosed early in follow-up is consistent with the strong association between infection and advanced-stage or lethal disease. For reverse causation to account for our study findings, the carcinogenic process would have to lead to higher levels of detectable antibodies. Although no data have been obtained to support or contest the assumption that levels of antibodies against T vaginalis increase during cancer development, tumorigenesis is known to alter adaptive immune response.[19] Second, our findings could be influenced by detection bias if men with T vaginalis infection were more likely to be diagnosed with prostate cancer. To address this possibility, we investigated the relationship of antibody levels to baseline PSA levels but found no association. However, we cannot rule out other urologic symptoms that could bring about diagnosis. Conservatively, serological history of infection with T vaginalis may be a marker of clinically relevant disease, as suggested by the association between infection and development of bony metastases or prostate cancer death. More research is required to establish this association.

Disease heterogeneity could also largely explain the apparent discrepancy between our findings and those of a recent study using data from 616 case subjects and 616 matched control subjects sampled from the Prostate Cancer Prevention Trial, a randomized trial of finasteride in 18 882 men, which found no association between T vaginalis seropositivity and the incidence of prostate cancer.[20] We found that a T vaginalis–seropositive status was principally associated with aggressive, potentially lethal disease. In contrast, most prostate cancers that were analyzed in the Prostate Cancer Prevention Trial were diagnosed at an early stage as a result of annual PSA screening and end-of-study prostate biopsy.[21] Evidence is accumulating that the risk factors for lethal and indolent prostate cancer may differ. In an analysis in the Health Professionals Follow-up Study that examined 10 risk factors for total or advanced prostate cancer supported by existing literature,[22] only four factors were found to have a statistically significant association with overall incidence: African American race, positive family history, higher tomato sauce intake (inversely), and α-linolenic acid intake. By contrast, recent smoking history, taller height, higher body mass index, positive family history, and high intakes of total energy, calcium, and linolenic acid were all statistically significantly associated with fatal prostate cancer. Consistent with our study, these results suggest that there may be multiple biological pathways that contribute to particular subgroups of prostate cancer.

The proportions of case subjects and control subjects with high seropositivity for antibodies against T vaginalis were somewhat higher in this study (24.5% of case subjects and 21.4% of control subjects) than in the Prostate Cancer Prevention Trial (15.2% of case subjects and 15.0% of control subjects) or the Health Professionals Follow-up Study (13% of case subjects and 9% of control subjects).[12] Assays for all three studies were prepared under the direction of the same microbiologist (J. F. Alderete) and used an enzyme-linked immunosorbent assay to detect antibodies against α-actinin protein from T vaginalis. In both the Prostate Cancer Prevention Trial and the Physicians' Health Study studies, known seropositive and seronegative control samples were used to determine absorbance score cut points, which were then applied to study case subjects and control subjects. In the Health Professionals Follow-up Study, absorbance score cut points were based on previous serological findings,[23,24] because serum samples from positive and negative control subjects were not available. Furthermore, absolute readings of the enzyme-linked immunosorbent assays in all three studies could be influenced by the specific technician conducting the assay and the fact that the laboratory was relocated in December 2007. Thus, differences in assay sensitivity may account for some of the variation in distribution of T vaginalis seropositivity across these three studies, especially given that demographic characteristics do not appear to explain the observed variability. All three studies included men from across the United States. Although African American race and lower socioeconomic status are generally associated with higher rates of sexually transmitted infections,[25] including T vaginalis infections,[3] the study with the highest proportion of men with a seropositive status (ie, the Physicians' Health Study) has the smallest proportion of African Americans (< 1%) and a relatively high socioeconomic status because all participants are physicians. Further, the mean age at blood collection in all three studies was similar (ie, 66 years in Health Professionals Follow-up Study, 64 years in Prostate Cancer Prevention Trial, and 59 years in Physicians' Health Study).

Because other sexually transmitted infections occur concurrently with T vaginalis infections, we cannot rule out the possibility that T vaginalis is acting as a marker for another infection. However, two studies[5,6] report that concomitant sexually transmitted infections, including those by N gonorrhoeae and C trachomatis, occur only in 10%–20% of case subjects, making it unlikely that these particular sexually transmitted infections could account for the observed association. Furthermore, the previous study in the Health Professionals Follow-up Study investigated other common sexually transmitted infections, including those by N gonorrhoeae, C trachomatis, Treponema pallidum, and human papillomavirus, and found no association with prostate cancer, except for an inverse association for human herpesvirus type 8 infection.[26,27] Nested case–control studies using data from the Nordic biobank consortium found no association between prostate cancer risk and human papillomavirus types 16, 18, and/or 33,[28] herpes simplex virus-2, or human herpesvirus type 8;[29] however, these studies observed a statistically significant inverse association with serological evidence of C trachomatis infection.[30] A study nested within the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial[20] found that seroprevalence of C trachomatis, human papillomavirus-16 and -18, herpes simplex virus-2, cytomegalovirus, and human herpesvirus type 8 were not individually associated with prostate cancer risk among white men. Men with one or more sexually transmitted infections, however, had a modest increase in risk of developing prostate cancer (OR = 1.3, 95% CI = 1.0 to 1.6), indicating that the measured infections could perhaps be serving as proxies for another infection such as T vaginalis.

Although our study may elucidate one mechanism by which local prostatic inflammation could arise and lead to downstream events that influence prostate cancer development and progression, studies that focus on local response to infection in the prostate are needed to determine whether T vaginalis is a causal agent. Nonetheless, in light of the limited understanding of factors that lead to lethal prostate cancer, our finding of an association between T vaginalis serostatus and aggressive prostate cancer is noteworthy. If our findings are confirmed, T vaginalis could serve as a marker for adverse outcomes in patients for prostate cancer or, more optimistically, as a target for secondary chemoprevention.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.