A pooled analysis of case-control studies of thyroid cancer - II. Menstrual and reproductive factors

Objective: It has been suggested that female hormones, and hence menstrual and reproductive factors, play a role in thyroid cancer etiology. Epidemiol ogical data, however, are limited and inconsistent, partly because of the s mall number of cases included in each study. To clarify the etiology of thy roid cancer, we conducted a pooled analysis of original data from 14 case-c ontrol studies, 4 from the United States, 2 from Asia, and 8 from Europe. Methods: This analysis included a total of 2,247 female cases of thyroid ca ncer (80% papillary) and 3,699 control women. Pooled odds ratios (OR) were estimated using logistic regression, conditioning on study and (i) matching sets for individually matched studies, or (ii) quinquennia of age for the other studies. Additional terms for age and history of radiation exposure w ere included in the regression equations. Results: The OR per year of later menarche was 1.04 (95% confidence interva l (CI) 1.0-1.1). Compared to premenopausal women, the OR was 1.3 for women with natural menopause, and 1.8 for those with artificial menopause, but th e studies were heterogeneous and the association may be due, at least in pa rt, to diagnostic or ascertainment bias. Parity, spontaneous or induced abo rtions and history of infertility were not associated with thyroid cancer r isk. The OR was above unity in women reporting later age at first birth (OR = 1.1, 95% CI 1.0-1.3 for 5-year delay) and higher in the first years afte r a birth. Conclusions: The associations of menstrual and reproductive factors with th yroid cancer risk were generally weak, but appeared stronger among women di agnosed with thyroid cancer at younger ages.