Differences in breast cancer risk factors according to the estrogen receptor level of the tumor

One hundred and forty-eight postmenopausal breast cancer cases and 585 postmenopausal controls were included in an investigation of whether various risk factors for breast cancer are associated with the level of estrogen (E) receptor (ER) protein in the tumor. In an intracase analysis, the tumor ER level was positively associated with nulliparity, late age at first live birth, a history of benign breast disease, and having breast-fed at least 1 child and was negatively associated with previous use of E replacement therapy. A case-control analysis suggested that the first three variables, established risk factors for breast cancer, are associated with an increased risk for malignant breast tumors that are ER-positive but not for those that are ER-negative. This analysis did not provide a clear interpretation of the findings in the intracase comparisons with regard to prior breast-feeding and the use of E replacement therapy.

The estimation of age, period and cohort effects for vital rates

In models for vital rates which include effects due to age, period and cohort, there is aliasing due to a linear dependence among these three factors. This dependence arises both when age and period intervals are equal and when they are not. One solution to the dependence is to set an arbitrary constraint on the parameters. Estimable functions of the parameters are invariant to the particular constraint applied. For evenly spaced intervals, deviations from linearity are estimable but only a linear function of the three slopes is estimable. When age and period intervals have different widths, further aliasing occurs. It is assumed that the number of deaths in the numerator of the rate equation has a Poisson distribution. The calculations are illustrated with data on mortality from prostate cancer among nonwhites in the U.S.

Cardiovascular disease and vasectomy. Findings from two epidemiologic studies

We performed two epidemiologic studies using routine abstracts of medical records to test the hypothesis that vasectomy may predispose men to cardiovascular disease. In a case-control study 1512 men who were under 55 years of age and had a history of myocardial infarction, stroke, or hypertension were matched with 3024 controls with other conditions; 2.4 per cent of the cases and 2.7 per cent of the controls were identified as having undergone vasectomy (risk ratio, 0.9; 95 per cent confidence limits, 0.6 to 1.3). In a cohort study data covering a mean period of 6 1/2 years after surgery were available on 1764 men who had had a vasectomy and on three comparison cohorts of men who had had other minor surgical procedures. There was no evidence of an increased risk of cardiovascular disease associated with vasectomy. There is no consistent evidence from our studies to support the hypothesis that in the short term vasectomy predisposes young men to cardiovascular disease.

Effects of cigarette smoking, alcohol, coffee and tea consumption on preterm delivery

A case-control study of 175 mothers of singleton, preterm infants and 313 mothers of singleton, term infants was undertaken at Yale-New Haven Hospital during 1977 to explore possible risk factors of preterm delivery. Heavy alcohol consumption (an average of two or more drinks per day) during the pregnancy was associated with an approximately 3-fold risk of preterm delivery. Women who smoked 10 or more cigarettes per day were also at an increased risk of a preterm delivery, but the effect of cigarette smoking was less pronounced once confounding variables had been taken into account. No association was observed between coffee drinking and shortened gestations. Tea consumption, especially four or more cups of tea per day, was more frequent among women with a preterm as compared to a term infant, but no significant relation was evident between heavy tea consumption and preterm delivery after controlling for the effects of other risk factors.

Covariance analysis for case-control studies with small blocks

Medical and epidemiological studies often employ a blocked design in order to control for some types of variation. For example, environmental and, to some extent, genetic factors may be controlled by blocking on sibs. The most common situation is one in which a subject in one group is blocked with one or more subjects in another. Additional factors may be considered as covariates in the analysis. Conditional likelihood methods give rise to consistent estimators of the parameters of the log-linear model. These estimators are generalized, and cases where there are small numbers of subjects in each group or where there are more than two groups are discussed. The resulting models can be considered as special cases of the general-linear model, and therefore the analyses can often be carried out with existing computer programs. The methods are illustrated by use of data from a case-control study in which each block contains one case and four controls.

A case-control study of cancer of the endometrium

A hospital-based case-control study of the epidemiology of endometrial cancer in women aged 45-74 years was carried out in Connecticut from 1977 to 1979. In total, 167 cases and 903 controls were included. Elevated risks were associated with the following factors: overweight, nulliparity, few pregnancies, use of estrogen replacement therapy, older age at menopause, and a history of ovarian or endometrial cancer in mother or a sister. Whites were more frequently affected than blacks, and better-educated women more often than less-educated women. Use of oral contraceptives was associated with a decreased risk, although the decrease did not reach statistical significance.

An epidemiologic study of hip fracture in postmenopausal women

In this case-control study of the epidemiology of hip fracture in post-menopausal women aged 45-74 years, cases of hip fracture and two control groups were selected from admissions to four general hospitals in Connecticut between September 1977 and May 1979. Fewer cases of hip fracture than controls had been exposed to estrogen replacement therapy, and among those who had been exposed, exposure time was shorter than that for controls. The cases had breastfed their children for shorter durations, and they more often had had both ovaries removed. Also, the cases were found to weigh less than the controls. The negative associations of hip fracture with estrogen replacement therapy, intact ovaries, and weight are consistent with the hypothesis that estrogens protect against hip fracture.

Effect of age at first childbirth on risk of developing specific histologic subtype of breast cancer

Epidemiologic variables related to breast cancer risk were assessed in a case-control study of 332 women with breast carcinoma and 1353 comparison women. Risk factors for breast cancer as a whole included nulliparity, late age at first childbirth, early age at menarche, late age at menopause, personal history of benign breast disease, family history of breast cancer, and among postmenopausal women, body weight. These risk factors were then analyzed with respect to histologic subtype of breast cancer involved, i.e., duct-derived or lobular tumors, to determine whether the association between any of the risk factors and breast cancer varied according to histopathologic subtype. Histologic subtype for the 316 cases reviewed included 284 duct cancers and 32 lobular carcinomas. Although slight differences were noted among some of the risk factors and the variety of cancer, none of the differences was marked except for the variable age at birth birth. For ductal carcinoma, the risk was highest among nulliparous women and decreased the younger a woman was at the time she gave birth to her first child. The risk of infiltrating lobular carcinoma, however, was lowest among nulliparous females or those who had given birth at a young age and increased the older a woman was when she gave birth to her first child.

Clinical characteristics and cigarette smoking in relation to prognosis of angina pectoris in Framingham

A long-term follow-up study of 130 Framingham, Massachusetts, men with newly diagnosed uncomplicated angina pectoris was undertaken in an attempt to identify medical and personal characteristics which could predict the likelihood of a future coronary event. Analyses indicated that, among the baseline characteristics, systolic blood pressure and electrocardiographic findings were the most powerful independent predictors of the outcome. Elevated blood pressure or an abnormal electrocardiogram increased coronary risk more than twofold. The lack of association between smoking at diagnosis and a new event appeared to be related to changes in habits after angina onset. Further scrutiny of the data indicated that, irrespective of blood pressure and electrocardiographic findings, those less than 60 years of age at angina onset who were nonsmokers or quitters during follow-up had a definite prognostic advantage over similarly aged continuing smokers. These results could not be explained by differences in coronary risk factors prior to symptom onset or by changes in factors other than smoking during follow-up. The findings suggest that stopping the cigarette smoking habit can improve both short-term and long-term prognosis in the younger patient and angina pectoris.

An epidemiologic study of epithelial carcinoma of the ovary

A case-control study to identify risk factors for epithelial ovarian cancer was undertaken among women in the age group 45-74 years who had been admitted to seven hospitals in Connecticut between July, 1977, and March, 1979. Characteristics that were found to increase the risk of epithelial ovarian cancer included being white, never having been pregnant, having a late age at menopause, having a family history of cancer of the ovary or endometrium, and having a long estimated number of years of ovulation. Prior use of post-menopausal estrogens did not alter the risk for epithelial ovarian cancer, but there was some indication that oral contraceptives protect against ovarian cancer. Women with ovarian cancer were somewhat more likely to have had a history of an underactive thyroid and were somewhat less likely to have had a history of an overactive thyroid than controls, although these trends were not statistically significant.

Exposure to prescribed drugs in pregnancy and association with congenital malformations

In a case-control (N = 1427 and 3001, respectively) study of deliveries in Connecticut the incidence of congenital malformations was 52 per 1000 live births. Of all mothers, 44.5% used at least 1 prescribed drug during pregnancy. Case mothers were more likely than controls to have used a prescription drug (odds ratio [o] = 13, P less than .0001), particularly an antidepressant (o = 7.6), narcotic analgesic (o = 3.6), or tranquilizer (o = 23); P less than .01 for all associations. There was a synergistic relationship with tranquilizer use and smoking in pregnancy, resulting in a o = 3.7 (P less than .01) risk for those exposed to both. The synergistic relationship of tranquilizers-cigarettes with malformations supports existing pharmacologic research and suggests that epidemiologic study of the impact of simultaneous maternofetal exposure to environmental agents may further explain the etiology of some congenital malformations.

Exogenous estrogens and other factors in the epidemiology of breast cancer

In a hospital-based case-control study of the epidemiology of breast cancer undertaken in Connecticut from 1977 to 1979, there was no evidence of an increase in risk for breast cancer among women who had used oral contraceptives or estrogen-replacement therapy. In fact, there was some suggestion of a decrease in risk for breast cancer with increasing length of use of oral contraceptives. Higher than average risks were found among women who had never give birth to a child, women with a late age at menopause, women, with an early age at menarche, women who had given birth to their first child at a relatively late age, women with previous benign breast disease, and women with a history of breast cancer in a sister or mother. Heavy women were at high risk for premenopausal breast cancer. The association between heaviness and postmenopausal breast cancer was strongest among women who had had their last menstrual period more than 5 years before the diagnosis of breast cancer.

Summary rates

The use of summary rates of reporting health data is discussed. Directly and indirectly adjusted rates are compared under two models of no interaction. A method of analysis is described whereby a model is first fitted to data in order to investigate assumptions, and subsequently the smoothed rates based on the model are summarized using an adjusted rate. We prefer directly adjusted rates when such a summarization is appropriate and the required data are available. An example is given which illustrates the method of estimation.

The analysis of rates and of survivorship using log-linear models

Models are considered in which the underlying rate at which events occur has a log-linear relationship with covariates. It is shown that the estimation of parameters involves the solution of identical systems of equations for data from either a Poisson process, an exponential distribution, a survival model or a generalized log-linear model. This enables one to use algorithms for fitting log-linear models, such as iterative proportional fitting (IPF), for the analysis of rates or survivorship.

Fibroadenoma in oral contraceptive users: a histopathologic evaluation of epithelial atypia

We have reviewed histologically a series of 120 fibroadenomas which formed part of the case material from a previous case-control epidemiological investigation of the relationship between oral contraceptive use and breast disease. We evaluated the epithelial component of the fibroadenoma for degree of cytologic atypia. This study indicates that the reduced risk for fibroadenoma among long-term users of oral contraceptives does not vary according to the degree of epithelial atypia present. This is in contrast to our previously reported findings for fibrocystic disease, in which the decreased frequency of occurrence of the disease in long-term users of oral contraceptives was found only for cases with no minimal epithelial atypia.

Induced abortion and congenital malformations in offspring of subsequent pregnancies

A case-control (n = 1427 and 3001, respectively) study in Connecticut found no relationship between delivery of an infant with congenital malformations and overall previous experience of induced abortion (odds ratio (o) = 0.9, 95% CL = 0.7, 1.1). Delivery of a congenitally malformed infant was also unrelated to: 1)abortion of penultimate pregnancy, 2) abortion of first pregnancy when index pregnancy is second, and 3) multiple previous abortions. Non-white women aged 25-29 who delivered a malformed child were significantly more likely to have aborted ( o = 2.6,95% CL = 1.2, 5.8, p less than 0.05). This may be due to more frequent histories of illegal and septic abortion, or to other characteristics of these women. Legal abortion performed under safe clinical conditions appears to impose no increased risk for subsequent delivery of a malformed infant.

Role of oral contraception in congenital malformations of offspring

In a case control study we examined the relationship between congenital malformations in offspring and maternal exposure to oral contraceptives around the time of conception. There were 1 370 with congenital malformations and 2 968 healthy control infants. Maternal oral contraceptive use was unrelated to malformations considered as a whole whether exposure last occurred in the year before conception [odds ratio = 0.9, p=0.25] or during pregnancy [odds ratio = 1.3, p=0.30]. Exposure during pregnancy doubled the risk for some specific diagnoses, including certain cardiovascular defects, but these increases were not statistically significant. Exposure to specific oestrogens or progestogens was also unrelated to the occurrence of malformations. There was a suggestion that women who both smoked more than 20 cigarettes per day and used oral contraceptives during pregnancy were more likely to deliver a malformed infant than were women who neither used oral contraceptives nor smoked during pregnancy.

The analysis of pair-matched case-control studies, a multivariate approach

In matched case-control studies one frequently must consider more than one variable in the analysis and in this paper a log-linear model is presented to meet this objective. A conditional argument yields a method for making inferences on the parameters measuring the association between the variables and disease. The result is similar to the problem of fitting a Bradley-Terry model when one has paired comparisons. Two methods of obtaining maximum conditional likelihood estimates of the parameters are possible: (i) fitting a quasi-independence model in the usual log-linear models context and (ii) fitting a linear logistic model. The results are illustrated with examples.

Fibrocystic breast disease in oral-contraceptive users. A histopathological evaluation of epithelial atypia

Epidemiological studies show a lower frequency of fibrocystic breast disease among long-term users of oral contraceptives than among women who have never used them. Fibrocystic disease may be a precursor of breast cancer; yet the incidence of breast cancer does not appear to differ between pill-takers and nontakers. To resolve this conflict, we examined the problem from a histologic standpoint in 205 premenopausal women, and found that this decreased frequency applied only to fibrocystic disease in which epithelial atypia was minimal or absent. In women with marked atypia there was no significant difference in frequency among long-term users as compared to women who have never used oral contraceptives. These findings suggest that a spectrum of cystic disease exists and that the long-term use of oral contraceptives protects against the forms of fibrocystic disease that are not firmly associated with an increased risk of breast cancer, but not against the premalignant forms.

Maternal smoking and congenital malformations: an epidemiological study

In a case-control study undertaken in several hospitals in Connecticut, it was found that women who reported smoking more than 20 cigarettes a day during pregnancy had a relative risk of about 1.6 for congenital malformations in the offspring of that pregnancy compared with women who said they had not smoked at all during pregnancy. However, there was no significant increase in risk among women who reported smoking 20 or fewer cigarettes a day during pregnancy compared with those who said they had not smoked at all during pregnancy. The higher risk among moderate and heavy smokers could not be attributed to any of the potentially confounding variables considered in this study; furthermore, it was specific to smoking during pregnancy rather than before pregnancy, and increased with the average amount smoked a day. Nevertheless, because the increase in risk was modest, because response bias could exist in a study of this type, and because no other studies have examined in detail the smoking-congenital malformation hypothesis, further research is needed to determine whether the relationship between maternal smoking and congenital malformations in offspring is causal.

Oral contraceptives and breast disease. An epidemiological study

A case-control study of the relationship between use of oral contraceptives and cancer of the breast, cancer of the corpus uteri, cancer of the ovary, and benign breast diseases was undertaken among women in the age group 20--44 years who had been admitted to several large hospitals in Connecticut. No significant association was seen between use of oral contraceptives and breast cancer. It was found that the longer a woman had used sequential oral contraceptives, the lower her relative risk of benign breast disease. Use of combination oral contraceptives, however, did not appear to influence the risk of benign breast diseases. The numbers of cases with cancer of the ovary and of the corpus uteri in this age group were too small for any conclusions to be reached.

Multivariate analysis for matched case-control studies

A multivariate method based on the linear logistic model is presented for the analysis of case-control studies with pairwise matching. This technique enables one to investigate the effect of several variables simultaneously in the analysis while allowing for the matched design. The odds ratio is used as the basic measure of risk. One is able to control for variables which are not matching variables while investigating the odds ratio for a particular factor, and to estimate the change in the odds ratio as the level of one or more interval variables changes. The computing methods used for obtaining maximum conditional likelihood estimates of the parameters of interest are modifications of standard programs for logit regression.

Life tables with concomitant information

A regression model for the analysis of survival data adjusting for concomitant information is developed. The model presented can lead to the log linear exponential model (Glasser [1967]) and the life table regression model of Cox [1972]. In addition, the model described can be used to analyze data from the commonly employed actuarial life table. A discussion of the special case where one is comparing two survival curves is presented. The methods developed are illustrated using data from a clinical trial investigating treatments for lung cancer.