Predictors for Thyroid Carcinoma in Israel: A National Cohort of 1,624,310 Adolescents Followed for up to 40 Years

Introduction: Data on adolescent precursors of thyroid cancer in adulthood are scant. Methods: In order to evaluate potential risk factors for thyroid cancer, we linked two national data sources: the military recruitment health examinations and the Israel National Cancer Register. The study population (1,624,310 participants) included 1,145,865 Jewish males aged 16-19 years when examined between 1967 and 2005, and 478,445 Jewish females aged 16-19 years when examined between 1989 and 2005. The cancer follow-up extended up to 2006. Multivariable Cox proportional hazards modeling was used. Results: During 24,389,502 person years of follow-up, 760 incidence cases of thyroid cancer were identified. The mean age at diagnosis was 25.2 +/- 4.2 years for women and 37.2 +/- 10.0 years for men. Women had a substantially higher incidence (birth cohort-adjusted hazard ratio (HR) = 5.70 [95% CI 4.45-7.31]; p < 0.001). Height predicted incidence in both sexes, with birth cohort-adjusted HRs of 1.03 (p < 0.001) in males and 1.04 (p < 0.001) in females, per 1 cm increment in height. In males, but not in females, there was a graded association between education, as measured by years of schooling, and incidence of thyroid cancer. Body mass index was not associated with incidence. In a multivariable analysis of 617,613 males and 469,185 females examined from 1989 onwards, which included sex, birth year, height, and education, the excess risk in females persisted strongly (HR = 5.67 [CI 4.30-7.13]), as did the association with height. Conclusions: Female sex, measured height in adolescence, and later birth cohorts were independent predictors of thyroid cancer in young and middle-aged adults in Israel. Further study is needed to unravel the mechanisms whereby height is associated with thyroid cancer.