Summary Plasma concentrations of certain hormones linked to breast cancer risk were measured in agepooled samples from 3,250 rural Chinese women in 65 counties, and 300 British women, all aged 35-64. In age-groups 35-44, 45-54 and 55-64 respectively, mean oestradiol concentrations were 36% (P = 0.043), 90% (P< 0.001) and 171% (P = 0.001) higher in the British than in the Chinese women, and mean testosterone concentrations were 48% (P<0.001), 68% (P<0.001) and 53% (P = 0.001) higher in the British than in the Chinese women. The difference in testosterone concentrations between the two countries appeared to be due largely to the lower average body weight in the Chinese women. Sex hormone binding globulin did not differ significantly between the two countries in age groups 35-44 and 45-54, but was 15% (P = 0.002) lower in the British than in the Chinese women at ages 55-64. Prolactin concentrations did not differ significantly between the two countries in any age group.The large international variation in breast cancer rates suggests that it may be possible to reduce the rates in high risk populations (Doll & Peto, 1981), but the reasons for this variation are still not fully understood. Breast cancer risk is clearly related to several reproductive factors (Kelsey, 1979). Pike et al. (1983) showed that the approximately six-fold higher rates of breast cancer in the USA than in Japan were not due to differences in age at first birth, nulliparity or age at menopause, but that differences in age at menarche and post-menopausal weight could explain some 70% of the difference in rates. They suggested that the remaining difference in breast cancer rates may be due to differences in hormone levels in the premenopausal period.The hypothesis that populations with low rates of breast cancer have low levels of oestrogens (and perhaps of other hormones) should be easily testable, but the results of previous studies which compared oestrogen levels in low and high risk populations have been confusing. In an early study, MacMahon and his colleagues found lower levels of urinary oestrone (El) and oestradiol (E2) in premenopausal Oriental women than in premenopausal Western women (MacMahon et al., 1974), but differences in such urinary oestrogens are not necessarily reflected in differences in plasma oestrogens. Moreover, Hayward et al. (1978) concluded from studies of a group of women in Tokyo that there were no striking differences in urinary El or E2 between Japanese and British premenopausal and post-menopausal women, and they reported a similar lack-of difference between the plasma oestrogens of these women. Gray et al. (1982)
Methods
Subjects: ChinaThe selection of normal subjects and collection of blood in China are described in detail in the monograph by Chen et al. (1990). Briefly, 65 rural counties throughout China were chosen to represent a wide range of mortality rates from seven of the most common cancers: cancers of the nasopharynx, oesophagus, stomach, liver, colorectum, lung and leukaemia. Two communes were c...