Plasma prolactin was measured in 2,572 pre-menopausal, 628 menopausal, and 1,666 peri- and post-menopausal women who were apparently healthy. Breast cancer was subsequently diagnosed in 47 of these women at a median time of 5 years after blood collection (pre-cancer cases). Prolactin levels in pre-menopausal cases increased significantly with age whereas this was not found in matched controls. The perimenopausal cases were characterized by extreme variability in prolactin levels. In post-menopausal women who developed breast cancer, the prolactin levels were significantly elevated, being at or above the 70th percentile for the controls. In this group the results are consistent with prolactin acting as a late-stage tumour promoter.

Normal premenopausal Japanese women have significantly more favourable mammographic parenchymal patterns (Wolfe Grades) than comparable British women. This finding is unaffected when the women are stratified by age, Quetelet's Index, age at menarche, age at first birth, and parity.

Breast parenchymal patterns have been assessed by the method of Wolfe in a prospective study in Guernsey. Women with P2 or DY grades had approximately four times the risk of developing breast cancer compared with those with N1 or P1 grades. Age, weight, parity and age at birth of first child are all related to the distribution of mammographic patterns. The combination of these variables with Wolfe grades may identify subsets of the population at very high risk.

Mammograms were obtained from 942 normal women aged over 30 years, and were classified by Wolfe's criteria. In women with an enhanced risk of breast cancer associated with a late age at first child or nulliparity, there was a significant increase in the proportion of P2 and DY mammographic patterns which are thought to be a radiological index of risk. There was a decrease in the proportion of P2 and DY grades in women using steroidal contraceptives. However, the association between radiological and epidemiological determinants of risk did not hold for age at menarche.

The three major urinary androgen metabolites, dehydroepiandrosterone (DHEA), androsterone (ANDRO), and aetiocholanolone (AETIO) were measured in 1,484 volunteer women between 29 and 60 years of age on the island of Guernsey from 1962-1967. Twelve of these women subsequently developed ovarian cancer after a median interval of 130 months and a minimum interval of 19 months. All three androgen levels in these women were lower than those in controls matched for age and menopausal status. The results were most striking for DHEA, where half of the cases were below the 27th percentile of their matched controls (p = 0.007, two-sided). The results for ANDRO were of marginal significance (p = 0.06), and those for AETIO were not significant (p = 0.33).