Undescended testis and the risk of testicular cancer: importance of sourceand classification of exposure information

Background The strength of the association between undescended testis and t esticular cancer varies considerably across studies. Here we report the eff ect of various classifications of self-reported history of undescended test is and different data sources on the estimates of the risk of testicular ca ncer from a case-control study. Methods We performed a population-based case-control study including 269 te sticular cancer cases and 797 controls matched on age and region. Medical h istory was assessed by interviews (index persons) and mailed questionnaires (mothers). We used conditional logistic regression to calculate odds ratio s (OR) and kappa coefficients to assess agreement between different sources of information. Results Odds ratios for testicular cancer ranged between 2.4 and 5.4 based on the sons' self-reports and between 1.1 and 1.9 based on the mothers' rep orts. The agreement between the sons and mothers on undescended, gliding or retractile testis was fair (kappa 0.53) and was good when these conditions were treated by surgery (kappa 0.89). The rating of a history of undescend ed testis by two urologists was fair (kappa 0.54). Conclusions The questionnaire design, the classifications of undescendcd te stis and data sources have an important impact on the OR for the associatio n of undescended testis and testicular cancer. These factors may partially explain the heterogeneity of the OR for this association in case-control st udies relying on self-reports.