Alice de M Zelmanowicz 1 , Mark Schiffman , Rolando Herrero , Alisa M Goldstein , Mark E Sherman , Robert D Burk , Patti Gravitt , Ray Viscidi , Peter Schwartz , Willard Barnes , Rodrigue Mortel , Steven G Silverberg , Julie Buckland , Allan Hildesheim
Sep 10 2005
Previous work suggests that cervical cancer may aggregate in families. We evaluated the association between a family history of gynecological tumors and risk of squamous cell and adenocarcinomas of the cervix in 2 studies conducted in Costa Rica and the United States. The Costa Rican study consisted of 2,073 women (85 diagnosed with CIN3 or cancer, 55 diagnosed with CIN2 and 1,933 controls) selected from a population-based study of 10,049 women. The U.S. study consisted of 570 women (124 with in situ or invasive adenocarcinomas, 139 with in situ or invasive squamous cell carcinomas of the cervix and 307 community-based controls) recruited as part of a multicentric case-control study in the eastern part of the United States. Information on family history of cervical and other cancers among first-degree relatives was ascertained via questionnaire. Information on other risk factors for cervical cancer was obtained via questionnaire. Human papillomavirus (HPV) exposure was assessed in both studies using broad spectrum HPV L1-based PCR testing of exfoliated cervicovaginal cells and in Costa Rica by additional testing of plasma collected from participants for antibodies against the L1 protein of HPV types 16, 18, 31 and 45 by ELISA. A family history of cervical cancer in a first-degree relative was associated with increased risk of squamous tumors in both studies (odds ration [OR] = 3.2 for CIN3/cancer vs. controls; 95% confidence interval [CI] = 1.1-9.4 in Costa Rica; OR = 2.6 for in situ/invasive squamous cell carcinoma cases vs. controls, 95% CI = 1.1-6.4 in the Eastern United States study). These associations were evident regardless of whether the affected relative was a mother, sister or daughter of the study participant. Furthermore, observed effects were not strongly modified by age. In Costa Rica, the effect persisted in analysis restricted to HPV-exposed individuals (OR = 3.0; 95% CI = 1.0-9.0), whereas in the Eastern United States study there was evidence of attenuation of risk in analysis of squamous carcinoma cases restricted to HPV positive women (OR = 1.4; 95% CI = 0.29-6.6). No significant association was observed between a family history of cervical cancer in a first-degree relative and adenocarcinomas (OR = 1.3; 95% CI = 0.43-3.9). History of gynecological tumors other than cervical cancer in a first-degree relative was not significantly associated with risk of disease in either study. These results are consistent with a role of host factors in the pathogenesis of squamous cell cervical cancer, although familial aggregation due to shared environmental exposures cannot be ruled out.