Associations seen in two separate cohorts of women A history of chlamydia infection doubled a woman's risk of developing ovarian cancer, according to a retrospective analysis of two different cohorts and control groups. The odds ratio for ovarian cancer ranged as high as 2.53 versus the control groups, depending on the cutpoint chosen for serum levels of chlamydia plasmid-encoded Pgp3, a serologic marker of chlamydia infection. The odds ratio for ovarian cancer ranged from 1.43 to 2.25 for the other cutpoints used to evaluate the two cohorts of women. Markers of other types of infection, including other sexually transmitted infections (STIs), did not have an association with ovarian cancer, Britton Trabert, MD, of the National Cancer Institute in Rockville, Maryland,, said at a press briefing prior to the American Association for Cancer Research (AACR) meeting. "Our findings support an association between chlamydia-associated pelvic inflammatory disease [PID] and ovarian cancer," Trabert said. "Null associations with other infections further support the specificity of the findings. If replicated, this study supports future studies to evaluate potential cancer risk reduction through treatment of chlamydia infections and PID." The findings have potential implications for earlier diagnosis and treatment of ovarian cancer, which could lead to better long-term outcomes, said Elaine Mardis, PhD, of Nationwide Children's Hospital in Columbus, Ohio, co-moderator of the AACR press briefing. "Ovarian cancer is typically diagnosed at quite late stage and has a poor prognosis," said Mardis. "This study focused on the evaluation of a common sexually transmitted disease, and the most important take-home point is that we might be able to go a long way in terms of preventing ovarian cancer through routine screening for agents, such as Chlamydia trachomatis." The study had its origin in evidence linking PID to ovarian cancer, although data have been mixed. For example, an analysis of 13 studies, published last year in the American Journal of Epidemiology, showed a significant association between PID and borderline tumors but not ovarian cancer overall. C. trachomatis is the leading cause of PID in developed nations, Trabert noted. However, ascertainment of PID or chlamydia infection in epidemiologic studies has proven challenging. To address the question of whether STIs are associated with ovarian cancer, investigators used a multiplex serology assay to measure serum antibodies to various infectious agents in two groups of women: 278 Polish women with ovarian cancer (plus an additional 278 women without the disease) and 160 women with ovarian cancer identified as part of the NCI-sponsored Prostate, Lung, Colorectal, and Ovarian (PLCO) screening trial (plus a control group of 159 women). The assay panel could detect the presence of Pgp3, as well as well as additional markers of chlamydia infection and serologic markers for other types of STI. Trabert's group performed statistical analyses using standard laboratory cutoff values for Pgp3, as well as two more stringent cutoffs identified in the Polish study and applied to the PLCO cohort. The results for the Polish cohort showed serum Pgp3 at the standard cutoff level was associated with an odds ratio for ovarian cancer of 1.63 (95% CI 1.20-2.22) compared with the control group, increasing to 2.00 with the second antibody cutoff (95% CI 1.38-2.89) and to 2.19 when the highest antibody titer threshold was applied (95% CI 1.29-3.73). Other markers of chlamydia infection also had significant associations with ovarian cancer when the higher cutoffs were applied, said Trabert. Analysis of data for the women from the PLCO showed a nonsignificant association with ovarian cancer with the standard laboratory cutoff value for Pgp3 was used (OR 1.43, 95% CI 0.78-2.63). When the first of the two more stringent cutoffs was applied, the association between Pgp3 and ovarian cancer achieved statistical significance (OR 2.25, 95% CI 1.07-4.71). The association did not reach statistical significance for the most stringent cutoff (OR 2.53, 95% CI 0.63-10.08). Markers for other infections did not have any statistically significant associations with ovarian cancer in either of the cohorts, including Mycoplasma genitalium, herpes simplex virus-1 or 2, human papillomavirus, hepatitis B or C, Epstein Barr virus, and cytomegalovirus. The AACR annual meeting takes place April 14-18 in Chicago. Source
