Also see:
Estrogen Levels Increase with Age
Estrogen Related to Loss of Fat Free Mass with Aging
Progesterone Decreases Aromatase Activity

Although we commonly think of the ovaries as the main source of estrogen, the enzyme which makes it can be found in all parts of the body. Surprisingly, in rhesus monkeys, aromatase in the arms accounts for a very large part of estrogen production. Fat and the skin are major sources of estrogen, especially in older people. -Ray Peat, PhD

J Clin Endocrinol Metab. 1985 Sep;61(3):564-70.
Fat tissue: a steroid reservoir and site of steroid metabolism.
Deslypere JP, Verdonck L, Vermeulen A.
Sex steroid concentrations and 17 beta-hydroxy-steroid dehydrogenase and aromatase activities were determined in fat tissue removed at surgery or, in order to allow comparisons in different sites, postmortem. Except for dehydroepiandrosterone (DHEA) sulfate (DHEAS), there existed a positive tissue/plasma gradient for all steroids studied (testosterone, androstenedione, DHEA, androstenediol, estrone, and estradiol), suggesting androgen uptake and estrogen synthesis in situ. Androgen concentrations did not vary according to site of origin of fat tissue, except that the DHEAS concentration was significantly lower in abdominal sc and omental fat than in breast, pericardial, or sc pubic fat. Tissue androgen concentrations were positively correlated with their plasma concentrations, but tissue and plasma estrogen concentrations were not correlated. All tissue steroid concentrations, with the exception of estradiol in men, decreased with age. Aromatase activity [androstenedione—-estrone; mean maximum velocity, 7.4 +/- 3.7 (+/- SD) fmol estrone/mg protein . h] did not vary between sexes or with site of origin of fat tissue. 17 beta-Hydroxysteroid dehydrogenase activity (estradiol—-estrone, mean maximum velocity 9.8 +/- 5.4 pmol/mg protein . h) was higher in fat from women than in that from men, higher in premenopausal than in postmenopausal women, and higher in omental than in sc fat. Its activity was noncompetitively inhibited in vitro by DHEA and DHEAS in near-physiological concentrations, and the enzyme activity was inversely correlated (P less than 0.001) with the tissue DHEA and DHEAS concentrations. We conclude that fat tissue is an important steroid hormone reservoir, that it is the site of active aromatase and 17 beta-hydroxysteroid dehydrogenase, and that tissue DHEA(S) may have a modulating effect on tissue estrogen production.

Clin Endocrinol (Oxf). 1978 Jul;9(1):59-66.
Sex hormone concentrations in post-menopausal women.
Vermeulen A, Verdonck L.
Plasma sex hormone concentrations (testosterone, (T), androstenedione (A), oestrone (E1) and oestradiol (E2) were measured in forty post-menopausal women more than 4 years post-normal menopause. Correlations between these and age, years post-menopause (YPM), degree of obesity and fat mass respectively were studied. T and A, as well as E1 and E2 were positively correlated (P less than 0.01), but no statistically significant correlation between A and E1 was observed. Sex hormone concentrations in this group of postmenopausal women (greater than 4YPM) did not show any variation as a function of age, with the possible exception of E2 which showed a tendency to decrease in the late post-menopause. E1 and to a lesser extent E2 as well as the E1/A ratio were significantly corelated with degree of obesity or fat mass, suggesting a possible role of fat tissue in the aromatization of androgens. Neither the T/A nor the E2/E1 ratios were correlated with fat mass, suggesting that the reduction of 17 oxo-group does not occur in fat tissue. The E1/A ratio was significantly higher than the reported conversion rate of A in E1. This might suggest the existence of an additional precursor of plasma E1.

Br J Obstet Gynaecol. 1985 Mar;92(3):260-5.
The effects of age and body composition on circulating serum oestrogens and androstenedione after the menopause.
Jensen J, Riis BJ, Hummer L, Christiansen C.
Circulating levels of oestrone, oestradiol and androstenedione were measured in two large groups of postmenopausal women, in one group the women were between 46 and 56 years of age and in the second, older group they were 70 years of age. In addition the fat mass was calculated from the height, weight and age of the women. Serum concentrations of both oestrogens did not change with age, whereas the serum androstenedione concentration decreased significantly. A change in body composition included decreased height and increased fat mass in the older group. Serum concentrations of both oestrogens correlated significantly with the fat mass and serum androstenedione as well as with each other. From the correlation analysis it may be concluded that the conversion rate of androstenedione to oestrone, and of oestrone to oestradiol, increases with age, which presumably explains the unchanged concentrations of the circulating oestrogens in relation to postmenopausal age, although the precursor decreases during the same period.