Positive correlation between parity and incidence of thyroid cancer: new evidence based on complete Norwegian birth cohorts.

We have examined the effect of parity on the incidence of thyroid cancer within a hazard model framework, using individual data on all (1.1 million) Norwegian women born 1935-1969. This unique data source was established by linking information from the Cancer Registry with maternity histories derived from the Central Population Register of Norway. The relative effects on the total incidence of thyroid cancer, net of age, place of residence and birth cohort, are 1.00 for the childless (arbitrarily chosen baseline group), 1.13 for parity 1, 1.30 for parity 2, 1.39 for parity 3 and 1.46 for parity 4+. The latter 3 figures are significantly higher than 1.00. The effects are larger for follicular carcinoma, and for this type of thyroid cancer there are also indications of a recent-pregnancy effect. Among women with 2 or more children, those who became pregnant less than 45 months previously have a higher cancer incidence than others at the same parity. Moreover, we have found that parity has virtually the same effect at all ages, in all birth cohorts and in both parts of the country. Age at first birth as well as the length of the interbirth interval appear to be unrelated to the thyroid-cancer risk. Our findings are consistent with previous case-control studies from other countries.