Is there visceral adipose tissue (VAT) intracellular hypercortisolism in human obesity?

The fact that obesity is a prominent feature of Cushing's syndrome (systemic hypercortisolism of adrenocortical origin) stimulated a 40-year search for evidence of systemic hypercortisolism in human obesity. That search has failed to find such evidence. For the past 15 years, however, studies have been done to evaluate a possible alternative type of hypercortisolism in obesity, namely visceral adipose tissue (VAT) intracellular hypercortisolism. The current review summarizes the evidence published so far about this possibility. There have been three types of evidence studied: direct measurement of the VAT levels of 11ß-hydroxysteroid dehydrogenase type I (11-HSD-1), which converts biologically inactive cortisone to biologically active cortisol; direct measurement of splanchnic cortisol production; and evaluation of the effect of a specific inhibitor of 11-HSD-1 on metabolic abnormalities associated with obesity, particularly diabetes mellitus. The results are complex and difficult to interpret. Our conclusion is that the presence of VAT intracellular hypercortisolism in human obesity is possible but unlikely.

Obesity and cortisol status.

The fact, that obesity is a prominent feature of hypercortisolism (Cushing's syndrome) has stimulated investigation on the possible existence of the reverse relationship, namely that hypercortisolism is a feature of obesity. We have reviewed half a century of literature on this question, and have found out the following: (1) Hypercortisolism can exist in two forms: systemic hypercortisolism, in which there is an overall bodily excess of cortisol, and tissue, or intracellular, hypercortisolism, in which there is increased intracellular concentration of cortisol without an overall bodily excess. (2) There are two parameters of systemic hypercortisolism: CPR and plasma cortisol concentration. Proper evaluation of the first parameter requires correction for the active metabolic mass, which is best performed by expressing CPR per gram of urinary creatinine. The second parameter can be confounded by the marked moment-to-moment fluctuations in plasma cortisol concentrations due to cortisol's episodic secretion. Proper evaluation requires measuring the 24-hour mean concentration. Of these two parameters of systemic cortisol status, the plasma concentration is the more critical and accurate. (3) Corrected CPR is normal in obese individuals, and 24-hour mean plasma cortisol concentrations are slightly but definitely subnormal. This combination of findings indicates diminished stimulability of the hypothalamic-pituitary-adrenal (HPA) axis, which normally regulates bodily cortisol status. This deduction is supported by empirical studies on HPA reactivity. (4) Tissue hypercortisolism, due to increased intracellular activity of 11beta-HSD-1, which catalyzes reduction of cortisone to cortisol, has been reported in obese mice and humans. The findings of various studies are not consistent, and whether the enzymatic overactivity is a cause or a result of obesity is still unclear.

Influence of postmenopausal hormone replacement therapy on an estrogen metabolite biomarker of risk for breast cancer.

Whether postmenopausal hormone-replacement therapy (HRT) increases the risk of breast cancer remains controversial, despite numerous epidemiological studies. We approached the question from a biochemical rather than an epidemiological direction - we hypothesized that if estrogen administration increases the risk of breast cancer, it should also alter a known estrogen biomarker of risk towards what has been observed in patients who already have breast cancer. The specific biomarker we studied was the ratio of the urinary excretion of two principal estradiol metabolites, 2-hydroxyestrone and 16 alpha-hydroxyestrone, which is markedly decreased in women with breast cancer and women with familial risk for breast cancer. We studied 34 healthy postmenopausal women not on HRT and 19 women on HRT (Premarin 0.625 mg daily plus Provera, 2.5 mg daily, in women with a uterus and Premarin alone in women without a uterus); treatment duration ranged from 3 months to 15 years. We also studied four women with recently diagnosed, untreated breast cancer. The women with breast cancer showed a significantly lower 2-hydroxyestrone to 16 alpha-hydroxyestrone ratio than control women on HRT (1.35 +/- 0.13 vs. 2.71 +/- 0.84; p < 0.0001). There was no significant difference in the metabolite ratio between healthy women on HRT and women not on HRT (2.82 +/- 0.92 vs. 2.71 +/- 0.84). There was no significant difference between women receiving Premarin alone and women receiving Premarin plus Provera (2.46 +/- 0.84 vs. 3.13 +/- 0.90), and neither differed significantly from women not on HRT (2.71 +/- 0.84). The finding that the ratio of women on HRT was not decreased to or toward the ratio in women with breast cancer can be interpreted, we believe, as a suggestive item of biochemical evidence that HRT is not a risk for breast cancer.

Sex difference in the effect of obesity on 24-hour mean serum gonadotropin levels.

To determine the effect of obesity on serum gonadotropin levels and any possible sex difference in the effect, we measured the 24-hour mean serum follicle-stimulating hormone (FSH) and luteinizing hormone (LH) concentrations in 62 healthy men with Body Mass Index (BMI) ranging from 20 - 94 and 61 healthy, regularly cycling women with BMIs ranging from 19 - 76. We also measured free testosterone (T) and estradiol (E2) in these subjects. There was a significant negative correlation between serum FSH and BMI in men: FSH(IU/L) = 49.9 x BMI -0.567; r = - 0.376, p = 0.0026; but a significant positive correlation between serum FSH and BMI in women: FSH(IU/L) =7.66 +/- 0.071 x BMI; r = 0.302, p = 0.018. Serum LH was weight-invariant in both sexes. In men, free T was negatively correlated with BMI: Free T (nmol/L) = 0.74 - 0.0068 x BMI; r = 0.585, p = 0.0381; and free E2 was positively correlated with BMI: Free E2 (pmol/L) = - 1.03 +/- 0.057 x BMI; r = 0.50, p = 0.0014. In obese women as a group, free T was higher than in lean women (33 +/- 6.8 S.E.M. vs. 17.4 +/- 2.0 pmol/L; p < 0.0001), and free E2 was also higher than in lean women: (6.90 +/- 0.80 vs. 4.84 +/- 0.55 pmol/L; p = 0.046). Of the many cases of hypothalamic-pituitary hormonal dysregulation that have been reported in obesity, none has been studied for sex differences. Our results mandate that possible sex differences be investigated in all cases of dysregulation.

Relationships of sex hormone levels to dependence in activities of daily living in the frail elderly.

OBJECTIVES: We undertook this nursing home study in order to determine the relationships between dependency in activities of daily living (ADL) and blood levels of estrone, testosterone, androstenedione, and dehydroepiandrosterone (DHEA). Little is known about this issue. METHODS: cross-sectional study of 370 nursing home residents. Hormone levels in blood specimens drawn in 1997 and 1998 were correlated with degree of ADL dependency recorded in medical charts. RESULTS: Because of multiple comparisons associations were deemed significant for P-values < or =0.017 for males and < or =0.0125 for females. In males, the following were inversely related: testosterone levels with dependency in transferring and eating; estrone with eating and a summary ADL index; and androstenedione with toileting and a summary ADL index (in all cases, r=-0.4; P=0.007-0.015). Inverse trends existed between testosterone levels and dependency in mobility and a summary ADL index; and androstenedione and eating (in all cases r=-0.3; P=0.030-0.055). Among females the following were directly related: estrone levels with dependence in mobility, toileting, transferring, and a summary ADL index; and DHEA with transferring and a summary ADL index (r=0.2-0.3, P=0.0001-0.01). Trends existed between estrone and eating, and DHEA and toileting (r=0.1-0.2, P=0.04). CONCLUSION: In male residents, higher sex hormone levels are associated with better ADL performance. Among females the opposite is true. While further studies are needed to elucidate these relationships, our results and recent findings of others suggest sex hormone actions in older women differ from those in younger populations. A possible stress-related mechanism is also presented.

Does postmenopausal estrogen administration increase the risk of breast cancer? Contributions of animal, biochemical, and clinical investigative studies to a resolution of the controversy.

Despite nearly six decades of epidemiological studies, meta-analyses, and reviews, there is still considerable controversy in the literature about the question, does postmenopausal estrogen administration increase the risk of breast cancer? In an effort to resolve the controversy, a number of animal, biochemical, and clinical investigative studies in this field have been reviewed. The following summary formulation is proposed: 1. Administration of estrogen is inherently capable of promoting the growth of breast cancer, and therefore of increasing the incidence of clinical breast cancer. 2. Human response to estrogen is like that of the low-cancer-incidence strains of mice studied by Lacassagne, in that large doses and prolonged administration are required to induce clinical breast cancer. 3. The blood levels of estradiol produced by the usual doses of postmenopausal estrogen are relatively low, equivalent to those of the follicular phase of the menstrual cycle. These levels may be near the threshold for producing breast-cancer-promoting effects; therefore, the tumor response will vary greatly in different populations, depending on genetic susceptibility factors: a. The prevalence of a family history of premenopausal breast cancer in a first-degree relative. b. The prevalence of abnormal BRCA1, BRCA2, and p53 genes. c. The prevalence of increased 16 alpha-hydroxylation of estradiol. d. The prevalence of smokers who are slow acetylators. 4. Consumption of alcohol (5 grams or more daily) along with the postmenopausal estrogen administration results in elevation of blood estradiol levels to values equivalent to those of the periovulatory peak of the menstrual cycle, which may be well above the threshold for producing breast-cancer-promoting effects in all women. The risk for cancer will therefore be uniformly increased in women who use alcohol and take estrogen. 5. Increased risk of breast cancer from postmenopausal estrogen administration can be eliminated by taking two synergistic steps: a. Eliminating alcohol consumption, or at least keeping it well below an average of 5 grams daily (equivalent to 2/3 ounce of whiskey or 3 ounces of wine). b. Diminishing the capacity to 16 alpha-hydroxylate estradiol, either through pharmacological agents such as indole-3-carbinol or through increased consumption of cruciferous vegetables. It is concluded that despite the inherent ability of postmenopausal estrogen therapy to increase the risk of breast cancer in theory, the increased risk can be eliminated in practice by minimizing or eliminating consumption of alcohol and ingesting pharmacological or dietary agents that reduce the 16 alpha-hydroxylation of estradiol.

Inpatient intervention in an indigent, minority population with uncontrolled diabetes.

OBJECTIVE: To study whether a program of brief, intensive, inpatient intervention could improve glycemic control in an indigent, minority population with uncontrolled diabetes unresponsive to outpatient treatment. METHODS: Patients with uncontrolled diabetes unresponsive to treatment in our outpatient Diabetes Clinic were admitted to our inpatient Diabetes Unit, where their care was directed by the Diabetes Team (an attending diabetologist, an endocrinology fellow, two nurses, and two nutritionists). Of 108 patients admitted, data were available for 96. Patients from minority populations constituted 91.7% of the group. All patients were indigent. The mean duration of stay was 4.3 days. After dismissal, patients underwent follow-up again in our Diabetes Clinic. During the 540-day follow-up period, 25 patients were electively readmitted when satisfactory improvement in glycemic control was not achieved. Hemoglobin A1c levels were averaged and plotted for the group at defined time points up to 360 days before admission and up to 540 days after admission. RESULTS: During the year before admission, hemoglobin A1c increased slowly from 10.1 +/- 0.3% (mean +/- standard error) at day -360 to 10.3 +/- 0.2% at day -210 (F5 = 29; P<0.01) and then rapidly to 11.4 +/- 0.2% at admission (F7 = 1,541; P<0.001). After admission, hemoglobin A1c declined rapidly to 9.5 +/- 0.2% at day 90 (F4 = 121; P<0.005), plateaued at that level until day 240, and then declined again slowly to 9.0 +/- 0.3% at day 540, the end of the follow-up period (F10 = 70; P<0.01). All hemoglobin A1c levels 30 days or more after admission were significantly lower than the mean level at admission (P<0.05 at day 30 and P<0.001 from day 45 to day 540). CONCLUSION: Brief, intensive, inpatient intervention in an indigent, minority population with uncontrolled diabetes unresponsive to outpatient treatment produced and sustained a significant improvement in glycemic control. This mode of treatment is a practical approach to achieving the improvement in glycemic control that the Diabetes Control and Complications Trial demonstrated to be effective in delaying the onset and slowing the progression of diabetic retinopathy, nephropathy, and neuropathy.

Can (99m)technetium methylene diphosphonate bone scans objectively document costochondritis?

STUDY OBJECTIVES: To determine whether bone imaging with 99mTc methylene diphosphonate is a specific method of making the diagnosis of costochondritis in patients with chest pain who rule out for myocardial infarction. DESIGN: Nonblinded prospective controlled study in 20 patients and 10 control subjects. SETTING: Inpatient medical service of a tertiary teaching hospital. PATIENTS: Two hundred consenting patients admitted to the hospital with chest pain and suspected myocardial infarction were examined. Those in whom acute myocardial infarction was ruled out were evaluated for the clinical signs of costochondritis, ie, tenderness over one or more costochondral junctions. Twenty patients who met the clinical criterion gave informed consent and were subjected to bone imaging. Ten control subjects with cancer who did not have clinical signs of costochondritis underwent bone imaging to rule out metastatic disease (normal in all cases). INTERVENTIONS: Bone imaging with I.V. 99mTc methylene diphosphonate. MEASUREMENTS: Bone scans of the investigative patients and the control subjects were read by two independent nuclear medicine specialists. RESULTS: Sixteen of the 20 patients with clinically diagnosed costochondritis showed increased technetium uptake at all costochondral junctions bilaterally; six of them also had increased uptake elsewhere on the chest wall (sternum, manubrium, or first rib). All 10 of the control patients likewise showed increased technetium uptake at all costochondral junctions bilaterally. CONCLUSIONS: Bone imaging with 99mTc methylene diphosphonate is not a specific method of making the diagnosis of costochondritis.

Impact of endocrine and diabetes team consultation on hospital length of stay for patients with diabetes.

PURPOSE: To determine whether consultation by an individual endocrinologist or by a multidisciplinary diabetes team (endocrinologist, diabetes nurse educator, and registered dietitian) can impact length of hospital stay of patients with diabetes. PATIENTS AND METHODS: Hospital stays of consecutive patients with a principal diagnosis of diabetes were compared. Forty-three patients were seen by an individual endocrine consultant and 27 were managed by the internist alone. Thirty-four patients were seen in consultation by the diabetes team. All consultations were performed at the request of the primary physician. There were no statistically significant differences among groups with respect to age, duration of diabetes, admitting diagnosis, glucose levels, or concomitant acute or chronic illness. RESULTS: Average length of stay of diabetes-team patients was 3.6 +/- 1.7 days, 56% shorter than the value, 8.2 +/- 6.2 days, of patients in the no-consultation group (P < 0.0001), and 35% shorter than the value, 5.5 +/- 3.4 days, of patients who received a traditional individual endocrine consultation (P < 0.05). The length of stay correlated with time from admission to consultation (regression equation: y = 3.92 + [1.09 x time to consultation]; r = .55; P < 0.0001). The slope (1.09) indicates that each 1-day delay in consultation resulted in a 1-day increase in length of stay. CONCLUSIONS: Length of stay was lowest in patients who received diabetes-team consultation. Three million Americans are hospitalized annually with diabetes at a cost of $65 billion. A team approach to their inpatient care may reduce their hospital stays, resulting in considerable health and economic benefits.

Testicular dysfunction in human immunodeficiency virus-infected men.

This review pertains to gonadal function in men with human immunodeficiency virus (HIV) infection, who often exhibit clinical and biochemical evidence of hypogonadism. Hypogonadotropic hypogonadism appears to be the most commonly encountered abnormality, although complete anterior pituitary insufficiency and primary gonadal failure have been reported. Levels of sex hormone-binding globulin (SHBG) are either unchanged or increased. Plasma levels of estrogens, progesterone, androstenedione, dehydroepiandrosterone sulfate (DHEA-S), and prolactin vary. Pathologically, except for involvement by opportunistic infections, no significant abnormality in the hypothalamic-pituitary area has been described, but evidence of orchitis is commonly present. The cause(s) of these abnormalities remains unclear. The possible factors leading to hypogonadism in HIV-infected men include HIV infection itself, opportunistic infections, chronic debilitating illness, and effects of cytokines on the hypothalamic-pituitary-gonadal axis. Further studies are needed to clarify the cause(s) of testicular dysfunction in HIV-infected men and its clinical significance, treatment, relevance to the progression of HIV infection, and influence on the immune system.

Twenty-four-hour mean plasma testosterone concentration declines with age in normal premenopausal women.

The 24-h mean plasma concentration of total testosterone (T) was measured in 33 healthy, regularly cycling, nonobese women between 21 and 51 yr of age. Percent free T was measured in 17 of them. Plasma dehydroepiandrosterone (DHEA) and dehydroepiandrosterone sulfate (DHEAS) were measured in 24 of them, and the DHEA-to-T and DHEAS-to-T ratios were calculated. It was found that the concentration of total T showed a steep decline with age; the regression equation was: T (nanomoles per L) = 37.8 x age-1.12 (r = -0.54; P < 0.003). According to this equation, the expected T concentration of a woman of 40 would be 0.61 nmol/L, about half that of a woman of 21 (1.3 nmol/L). The percent free T did not vary significantly with age, so free T concentration likewise showed a steep decline with age. The DHEA-to-T and DHEAS-to-T ratios were both age invariant, clearly because the levels of DHEA and DHEAS also decline steeply with age, as previously reported.

Hormonal profiles in women with breast cancer.

The literature findings on endogenous hormonal profiles in women with breast cancer are reviewed in detail. It is concluded that four sets of findings are valid: (1) diminished adrenal androgen production, probably genetic, in women with premenopausal breast cancer; (2) ovarian dysfunction (luteal inadequacy plus increased testosterone production) in breast cancer at all ages; (3) increased 16 alpha-hydroxylation of estradiol in breast cancer at all ages; and (4) evidence that prolactin is a permissive risk factor for breast cancer, and that the pregnancy-induced decrease in prolactin levels may account for the protective effect of early pregnancy against breast cancer.

The relationship between serum levels of insulin and sex hormone-binding globulin in men: the effect of weight loss.

It is known that there is an inverse relationship between the serum levels of insulin and sex hormone-binding globulin (SHBG) in women, but the relationship in men has not been reported. It is not known whether changes in the one cause changes in the other, or whether they change in opposite directions in response to some third factor. Because obesity raises insulin levels and lowers SHBG levels in both sexes, we proposed to study the cause-effect question by determining whether the relationship between changes in SHBG and insulin levels during active weight loss. We studied 70 healthy weight-stable men with body mass index (BMI) from 20.7-94 (normal, 22.5 +/- 2.5) and restudied 17 of them during diet-induced weight loss. Fasting serum insulin levels in the weight-stable men showed a positive linear correlation with BMI, increasing 1 microU/mL per unit increase in BMI (P < 0.0001). SHBG levels in the weight-stable men showed a negative linear correlation with BMI, decreasing 0.2 nmol/L per unit increase in BMI (P < 0.0002). In the weight-stable men, there was an inverse hyperbolic correlation between SHBG and insulin levels; SHBG (nmol/L) = 13.1 + [30.1 divided by insulin (microU/mL)] (P < 0.002). During weight loss, insulin levels decreased at an average rate of 6.1 microU/mL per unit decrease in BMI, a much higher slope than the positive slope vs. BMI in weight stable men. During weight loss, SHBG levels increased at an average slope of 0.43 nmol/L per unit decrease in BMI, much higher than the negative slope of 0.2 nmol/L per unit increase in BMI in weight-stable men. Values for the SHBG vs. insulin coordinates in the weight-losing subjects did not differ significantly from those expected from the SHBG vs. insulin equation in weight-stable subjects. The stability of the SHBG-insulin relationship during weight loss despite the profoundly altered relationship of each separate component to BMI strongly suggests a close metabolic link between SHBG and insulin. As SHBG is not known to alter the production or metabolism of insulin, whereas insulin has been shown in vitro to decrease the synthesis of SHBG, it seems a reasonable conclusion that the predictable inverse relationship between serum insulin and SHBG indicates that insulin controls SHBG synthesis in vivo.

Abnormal production of androgens in women with breast cancer.

Two long and broad streams of medical literature, from the 1950's to date, have established the existence of two unrelated abnormalities of androgen production in women with breast cancer. One is the genetically determined presence of subnormal production of adrenal androgens (i.e. DHEA and DHEAS) in women with premenopausal breast cancer and their sisters, who are at increased risk for breast cancer. The other is excessive production of testosterone, of ovarian origin, in subsets of women with either premenopausal or postmenopausal breast cancer and women with atypical breast-duct hyperplasia, who are at increased risk for breast cancer; along with the hypertestosteronism, there is frequently chronic anovulation in the premenopausal patients. The combination of ovarian hypertestosteronism and chronic anovulation is characteristic of the polycystic ovary syndrome and is also frequently seen in women with abdominal ("android") obesity; both PCOS and abdominal obesity are known to be characterized by high risk for postmenopausal cancer. The elevated testosterone levels and the increased levels of insulin, IGF-I, and IGF-II that are seen in PCOS and abdominal obesity could favor the development of breast cancer in several ways, all of which have been demonstrated experimentally: binding of testosterone to cancer cells bearing testosterone receptors, with direct stimulation; intratissular aromatization of testosterone to estradiol, with stimulation of estrogen-sensitive cells; stimulation of the production of epithelial growth factor (EGF) by testosterone, with direct mitogenic effect of EGF on cancer cells; stimulation of aromatase by insulin and IGF-I; direct mitogenic stimulation of cancer cells by insulin, IGF-I, and IGF-II; and stimulation by IGF-I and IGF-II of the intratissular reduction of estrone to estradiol. Since PCOS is probably largely genetically determined, and abdominal obesity may also be, the hypertestosteronism of these conditions may represent a second genetically determined hormonal risk factor for breast cancer.

A randomized double-blind crossover study of the antiobesity effects of etiocholanedione.

Etiocholanedione (ED), a natural metabolite of dehydroepiandrosterone, has antiobesity effects in animals when given orally and is nontoxic. We carried out a trial of oral ED in obese humans. In a 20-week randomized double-blind crossover study, 14 subjects lost significantly more weight and body fat during treatment with oral ED, 4 gm daily, than during placebo administration. Mean weight loss during ED administration was 2.8 +/- 5.5 kilograms, which was equivalent to 0.53 +/- 0.91 kilograms per week per 100 kilograms of body fat; mean weight change during placebo administration was essentially zero: +0.21 +/- 4.2 kg, or +0.04 +/- 0.74 kg/wk/100 kg body fat. The difference between the weight changes in the two periods was significant: for delta kg, P < 0.05; for delta kg/wk/100 kg body fat, P < 0.03. Densitometric measurement of body fat content showed that the mean weight loss coincided almost exactly with the mean decrease in fat content; thus, over the 10-week period of ED administration, the mean fat loss was about 5% of the initial body fat content. Three of the obese subjects had strikingly greater fat loss, about 18%, 19%, and 25% of the initial body fat content. There were no significant subjective or objective side effects of ED administration.

A perspective on the hormonal abnormalities of obesity: are they cause or effect?

Studies in our laboratory and elsewhere have demonstrated numerous abnormalities of steroid and polypeptide hormone secretion in obesity: hyperestrogenemia and hypogonadotropic hypogonadism in obese men; diminished SHBG levels in both sexes; elevated free testosterone and free estradiol in obese women; PCOS-like gonadotropin and sex-hormone abnormalities in obese women; elevated serum insulin in both sexes; blunted stimulability of prolactin, growth hormone, and vasopressin in both sexes; and elevated basal levels and blunted stimulability and suppressibility of beta-endorphin in both sexes. All of these abnormalities have been clearly shown to be partly or completely reversible with weight loss, with the exception of the endorphin abnormalities. In that area, four out of the five studies reported show no reversibility with weight loss. Reversibility of nearly all the hormonal abnormalities of obesity (i.e., all but the hyperendorphinemia) by weight loss suggests that none of them is causative of obesity. Nevertheless, some of the reversible abnormalities may secondarily amplify the morbidity associated with obesity: the hyperinsulinemia may be related to the increased risk of hypertension, hyperlipidemia, coronary disease, and Type II diabetes; the elevated levels of free estradiol in obese women may be related to their increased risk of breast and endometrial cancer. The role of hyperendorphinemia in obesity clearly requires further investigation, since it is the only observed hormonal abnormality that appears to be non-reversible by weight loss, and also since there seems to be increased sensitivity to beta-endorphin in obesity. The possibility that endorphin abnormalities may be causal in obesity cannot be ruled out.

Biological and endocrinological insights into the possible breast cancer risk from menopausal estrogen replacement therapy.

The question of whether estrogen therapy increases the risk of breast cancer is reviewed. Despite more than 60 epidemiological studies and several meta-analyses over a five-decade period, there is no consensus about the answer. At present, the majority of investigators agree that short-term or medium-term therapy (less than 10 years) poses no measurable risk; some, but not all, investigators feel that there is a modest risk with long-term therapy (more than 15 years). Even this semi-consensus is clouded by the startling and clear-cut finding of the largest ever epidemiological study, the Nurses Surveillance Study, that a small increase in risk with estrogen therapy occurred only in women who also ingested alcohol, itself a known risk factor for breast cancer; women who did not ingest alcohol were at no increased risk. Because virtually none of the other epidemiological studies has controlled for alcohol ingestion, the conclusions of all of them are placed in doubt. To try to shed light on this problem, the 60-year-old studies of Lacassagne et al. on the induction of breast cancer in mice by estrogens were reviewed. They found that the magnitude and timing of the inducing effect of estrogen depended on the spontaneous breast cancer incidence in the mouse strain studied: in no-incidence strains, no cancer was induced; in high-incidence strains, induction was rapid and universal; in low-incidence strains, only a low percentage of animals had cancer induced, and it required prolonged estrogen administration.(ABSTRACT TRUNCATED AT 250 WORDS)