Search

Ages at menarche and first birth are established risk factors for breast cancer. The interval between these ages may also affect risk, since the breast is more susceptible to carcinogenic insults during this period than during the parous period. However, few investigators have studied this relation. Using logistic regression, the authors evaluated associations between the timing of reproductive events and breast cancer risk among 4,013 cases and 4,069 controls enrolled in a multicenter, population-based US case-control study of White and African-American women (1994-1998). For White, parous premenopausal and postmenopausal women, those who had an interval of > or =16 years between the ages of menarche and first birth had 1.5-fold (95% confidence interval (CI): 1.0, 2.2) and 1.4-fold (95% CI: 1.1, 1.8) increased risks of breast cancer, respectively, in comparison with those who had < or =5 years between these ages. Adjusting for age at first birth altered these risk estimates somewhat, to odds ratios of 1.5 (95% CI: 0.8, 2.9) and 1.0 (95% CI: 0.6, 1.5), respectively. These associations were stronger for lobular and hormone-receptor-positive tumors but were absent among premenopausal African-American women. The authors conclude that the interval between age at menarche and age at first birth is associated with the risk of hormonally sensitive types of breast cancer, particularly among White women.

Smoking is one of the few risk factors that have been identified for functional ovarian cysts, and results of one epidemiologic study suggest that body mass index (BMI; weight (kg)/height (m)(2)) may modify the effect of this exposure. The current study assessed the association of cigarette smoking and marijuana use with functional ovarian cyst risk by using data from a population-based 1990-1995 case-control study of 586 incident functional ovarian cyst cases and 757 age-matched controls in a large health maintenance organization in Washington State. In multivariate analyses controlling for age, education, and reference year, the authors found an increase in risk associated with current cigarette smoking among women whose BMI was20 (odds ratio (OR) = 2.48, 95% confidence interval (CI): 1.32, 4.64) or 20-25 (OR = 1.60, 95% CI: 1.04, 2.46) but not25 (OR = 0.85, 95% CI: 0.53, 1.37). Corresponding risks associated with current marijuana use were BMI20, OR = 2.05 (95% CI: 0.89, 4.75); BMI 20-25, OR = 1.78 (95% CI: 1.00, 3.17); and BMI25, OR = 0.72 (95% CI: 0.36, 1.42). Study results indicate that increased BMI may attenuate the adverse effect of smoking on the risk of functional ovarian cyst.

A validation study evaluated the accuracy of self-reported use of commonly used medications among older women. Within a case-control study of breast cancer, drug information was ascertained by interview. Pharmacy records from 1990 to 1999 were obtained from a Washington State health maintenance organization (66% of subjects) and retail pharmacies (34% of subjects) on a sample of subjects (212 cases, 191 controls) and used as the "gold standard." Recall accuracy was assessed for 6-month, 2-year, and 8-year time windows. Sensitivity of antihypertensive use was 92% (95% confidence interval (CI): 85, 96) for cases and controls in the 6-month period and slightly lower for the 2-year (90% (95% CI: 82, 94) and 87% (95% CI: 78, 92)) and 8-year (80% (95% CI: 69, 88) and 79% (95% CI: 68, 88)) periods. For statins, sensitivity was 83% (95% CI: 64, 93) for cases and 93% (95% CI: 69, 99) for controls in the 6-month period, 75% (95% CI: 55, 88) and 86% (95% CI: 60, 96) in the 2-year period, and 67% (95% CI: 42, 85) and 75% (95% CI: 41, 93) in the 8-year period. For self-report of antidepressants, sensitivities ranged from 66% (95% CI: 47, 80) in the 6-month period to 44% (95% CI: 30, 60) in the 8-year period. Specificity was high among all drug classes, ranging from 91% to 100%. Recall did not differ by case-control status. Trivial changes in estimates were observed when health maintenance organization records alone were used as the gold standard. Self-reported use of antihypertensives and statins appears to be relatively accurate among older women.

The authors investigated the possibility that, in interview-based case-control studies, controls are more likely than cases to underreport a history of induced abortion. A case-control study was conducted in White women under 45 years of age who had given birth in Washington State during 1984-1994. The cases were women in three metropolitan counties of Washington State diagnosed with invasive breast cancer during 1984-1994; controls were selected through random digit dialing. A history of induced abortion among study participants was compared between interview data and information collected on the birth record of the last child to whom they gave birth (225 cases, 303 controls). Among women with a prior induced abortion recorded on the birth record, 14.0% of the 43 cases and 14.9% of the 47 controls did not report an induced abortion at interview (difference = -0.9%, 95% confidence interval of the difference: -15, 14). The authors' data do not suggest that controls are more reluctant to report a history of induced abortion than are women with breast cancer.

To evaluate whether recreational physical activity is associated with breast cancer among young women, the authors analyzed data from a population-based case-control study. Cases (n = 1,668) were women under age 45 years who had been newly diagnosed with breast cancer between 1990 and 1992 in Atlanta, Georgia, central New Jersey, or Seattle, Washington. Controls (n = 1,505) were frequency-matched to cases by 5-year age group and geographic area of residence. Breast cancer was not associated with recreational activity in any of the three time periods assessed (highest quartile of activity vs. lowest: age- and center-adjusted odds ratio (OR) = 0.94 (95% confidence interval (CI) 0.77-1.15) at ages 12-13 years, OR = 1.08 (95% CI 0.88-1.32) at age 20 years, and OR = 1.18 (95% CI 0.97-1.44) during the past year), with the average of the three time periods (OR = 1.02, 95% CI 0.84-1.25), or with daily climbing of at least two flights of stairs (without stopping) during the past year (daily climbing vs. never climbing: OR = 1.03, 95% CI 0.86-1.23). Estimates were not modified or confounded by body mass index, menopausal status, or caloric intake during the past year. These results do not support a protective role for physical activity in the risk of breast cancer among young women.

The authors used data from a population-based case-control study of breast cancer in women aged < or = 44 years (cases, n = 975; controls, n = 866) conducted between 1994 and 1996 in three counties of western Washington state to assess the validity and reliability of reported perinatal factors. For a sample of participants, exposure information from self-administered questionnaires was validated with information from birth certificates (cases, n = 378; controls, n = 283). Detailed information regarding perinatal characteristics of their daughters was also collected from subjects' mothers (case mothers, n = 510; control mothers, n = 436) to assess the reliability of subjects' reporting of these events. Although reporting of birth weight by subjects (cases, r = 0.83; controls, r = 0.80) and their mothers (case mothers, r = 0.89; control mothers, r = 0.84) was highly correlated with the birth certificates, there was differential measurement error by subjects; cases reported birth weight accurately on average, but controls tended to underestimate their birth weight. Agreement between the subject and mother report was excellent for birth weight (cases, r = 0.85; controls, r = 0.87) and good for other perinatal factors, but birth order and maternal diethylstilbestrol use were underreported among cases and reported accurately among controls. Differential measurement error of birth weight by case-control status resulted in biased odds ratios for breast cancer risk.

A population-based case-control study was conducted to assess the relation between physical activity and colon cancer among men and women aged 30-62 years. Cases were 251 men and 193 women diagnosed with colon cancer in 1985-1989 in three countries in the Seattle metropolitan area who were identified from the Seattle-Puget Sound Surveillance, Epidemiology, and End Results registry. Controls were 233 men 194 women identified by random digit telephone dialing who were selected by stratified random sampling to approximate the age, sex, and county distribution of cases. Physical activity was assessed by questions on frequency and duration of types of recreational and occupational activities during the 10-year period ending 2 years before diagnosis. Each activity was classified as low intensity (4.5 METs ) or moderate to high intensity (or = 4.5 METs). For men and women combined, moderate or high intensity recreational activity was associated with a decreased risk of colon cancer (relative risk (RR) for two or more times per week vs. none = 0.70, 95% confidence interval (CI) 0.49-1.00). This relation was stronger for men than women. Occupational activity was not associated with colon cancer, except among men younger than 55 (RR foror = 14.5 hours per week of moderate activity vs. none = 0.29, 95% CI 0.12-0.69). Among men and women combined, total moderate or high intensity activity (occupational plus recreational) was marginally related to colon cancer (RR foror = 5 hours per week vs. none = 0.78, 95% CI 0.55-1.10). These results were adjusted for age (and sex in the combined sex analyses) and were not confounded by body mass index, dietary factors, or other measured health behaviors. The results of this study provide modest support to the growing number of studies showing that recreational and/or occupational physical activity is associated with a reduced risk of colon cancer.

This population-based case-control study assessed the effect of current cigarette smoking on the risk of functional ovarian cyst development. Cases were all 15-39 year-old enrollees of Group Health Cooperative of Puget Sound (GHC) who had either an inpatient primary diagnosis of functional ovarian cyst in 1988 or 1989 (n = 61) or an outpatient primary diagnosis of functional ovarian cyst at five GHC primary care clinics (n = 37). Controls were randomly selected enrollees matched to cases on age, primary care clinic, and enrollment date (n = 239). A total of 36.7% of cases and 20.5% of controls were determined by medical record review to be current smokers. Compared with women who were not current smokers, the relative risk of a diagnosed functional ovarian cyst among smokers was 2.0 (95% confidence interval 1.1-3.5). This finding, in an area of little previous research, provides a promising lead in the investigation of the etiology of functional ovarian cysts and the effects of smoking on ovarian function in general.

A population-based case-control study was conducted in western Washington state to examine the relations between infection with human papilloma viruses (HPV), herpes simplex viruses (HSV), and risk of oral squamous cell cancer in men. Interviews were completed on 131 oral cancer cases diagnosed between January 1985 and December 1989 and 136 controls frequency matched to cases on age and date of diagnosis who were obtained by random digit dialing. The risk for oral cancer among men with 30 or more sexual partners was 2.4 times that of men with four or fewer partners (95% confidence interval (CI) 1.0-5.9). Men who ever practiced oral sex had lower risk for oral cancer relative to men who never practiced oral sex (relative risk (RR) = 0.4, 95% CI 0.2-0.8). Analyses of exfoliated oral cavity cells for the presence of HPV-6 DNA with polymerase chain reaction revealed that men with an oral HPV-6 infection had 2.9 times the risk for oral cancer of noninfected men (95% CI 1.1-7.3), whereas men with an oral HPV-16 infection had 6.2 times the risk for oral cancer of noninfected men (crude RR = 6.2, 95% CI 0.7-52.2). Relative risks associated with serologically detected HSV-1 and HSV-2 infections were 0.8 (95% CI 0.3-1.7) and 1.8 (95% CI 0.7-4.6), respectively. The authors conclude that HPV-6 is associated with oral cancer. Although men infected with HPV-16 and HSV-2 were at elevated risk, these associations may have been due to chance. The role of specific sexual practices in the transmission of viruses to the oral cavity remains unclear.

Data from a population-based case-control study of chronic lymphocytic leukemia were analyzed to assess the possible etiologic role of chronic antigenic stimulation. The study, conducted in four geographic areas of the United States (the metropolitan areas surrounding Seattle, Washington, Salt Lake City, Utah, Detroit, Michigan, and Atlanta, Georgia) sought to identify all incident cases (n = 430) among residents diagnosed between July 1, 1977 and December 31, 1981. The responses of these cases to questions about possible sources of antigenic stimulation were compared with the responses of controls selected from the populations of these areas. Little difference between cases and controls was present for a history of most forms of viral and bacterial infection and for a history of allergies or allergy treatment. However, a relation was observed with antecedent syphilis infection (odds ratio (OR) = 5.0, 95% confidence interval (Cl) 2.0-12.9). Associations of smaller magnitude were observed with a history of tuberculosis (OR = 1.9, 95% Cl 1.0-3.7) and of urinary tract infection (OR = 1.4, 95% Cl 1.1-1.9). Overall, however, the authors found little evidence of a relation between chronic antigenic stimulation and the occurrence of chronic lymphocytic leukemia. Nonetheless, because the measure of prior antigenic stimulation was restricted to that obtained through interviews and undoubtedly was an insensitive one, these negative results should not be interpreted as ruling out antigenic stimulation as a possible cause of some cases of chronic lymphocytic leukemia.

Electric blankets are an important domestic source of electromagnetic fields (EMF) because of the relatively high intensity of emission, prolonged exposure, and intimate contact with the source. In a case-control study of testicular cancer in western Washington during 1981 to 1984, the relation between EMF exposure from electric blankets and the occurrence of testicular cancer was examined. The respective proportions of cases and controls who reported the use of an electric blanket were almost identical (age-adjusted rate ratio (RR) = 1.0, 95% confidence interval (CI) 0.7-1.4). Distributions of the duration of use were also very similar in cases and controls. Compared with controls, the frequency of use of an electric blanket was slightly lower in men with seminoma (RR = 0.7, 95% CI 0.5-1.2) and slightly higher among men with nonseminoma germ cell tumors (RR = 1.4, 95% CI 0.9-2.3). Overall, the results of this study suggest that increased exposure to EMF from electric blankets contributes little, if at all, to the risk of testicular cancer in adult white men.

When conducting epidemiologic case-control studies, some investigators include only controls who can be interviewed within a certain time after contact and/or do not recontact potential participants who initially refuse, whereas others expend considerable effort to recruit reluctant respondents. This additional effort is only worthwhile if it results in a sample that is more representative of the target population. In this study, the authors used data collected from in-person interviews of 5,616 female controls to compare characteristics of willing, accessible respondents with those of their less accessible or less willing counterparts to determine whether or not the two groups differed with respect to lifestyle, socioeconomic status, health history, and demographic characteristics. Late responders were younger, were more likely to be non-White, were less likely to have attended college, and were more likely to be current smokers than early responders. Initial refusers were similar to late responders with respect to education and race. Initial refusers were also older, were less likely to be currently married, were less likely to have a managerial occupation, had fewer lifetime sexual partners, and were more likely to have a history of diabetes than early responders. These findings suggest that additional effort expended in recruiting reluctant respondents may often result in more accurate estimates of population characteristics that are of interest in epidemiologic research.

To gain insight into whether breast cancer tumors jointly classified by estrogen receptor (ER) and progesterone receptor (PR) status represent diseases with differing etiologies, data from a population-based case-control study of US women 20-44 years of age were analyzed. Cases included 1,556 women diagnosed between 1990 and 1992. Age- and geographic-frequency-matched controls included 1,397 women identified by random digit dialing. Heterogeneity between ER+PR+ and ER-PR- tumors was most pronounced in relation to age, race, and recreational exercise at 12-13 years of age. Multivariate-adjusted odds ratios for ER+PR+ tumors were 0.64 (95% confidence interval (CI): 0.47, 0.89) for 30-34 versus 40-44 years of age, 0.89 (95% CI: 0.63, 1.25) for Black versus White race, and 0.84 (95% CI: 0.68, 1.03) for exercise at 12-13 years of age above versus at or below the median. Corresponding odds ratios for ER-PR- tumors were 1.24 (95% CI: 0.86, 1.77), 1.51 (95% CI: 1.07, 2.14), and 1.15 (95% CI: 0.90, 1.48). Risk of ER-PR- cancer in relation to menstrual and reproductive (parity and lactation) characteristics, alcohol consumption, and family history of breast cancer was similar to that observed for ER+PR+ tumors. These findings only modestly support the hypothesis that hormonally related risk factors have differing relations with ER+PR+ versus ER-PR- tumors among younger women.

The authors conducted a population-based case-control study of 410 women residing in three counties in western Washington State who were aged 18-64 years when diagnosed with papillary thyroid cancer in 1988-1994 and 574 controls to assess the effects of pregnancy history and other aspects of reproductive life on risk of this disease. Among women aged 45-64, the authors observed no associations with number of live births, age at first live birth, or age at last live birth. Risk was somewhat increased in women45 years who had given birth within the previous 5 years; this association was most evident among women who reported that cancer symptoms had led to diagnosis. Among women who had given birth within the last 5 years, risk was greatest among those with two or more births during that time period (relative risk (RR) = 4.2, 95% confidence interval (CI): 2.0, 8.9, relative to parous women whose last birth was5 years before the reference date). Risk of thyroid cancer was also associated with lactation during the previous 5 years (e.g., RR = 2.9, 95% CI: 1.5, 5.5, among parous women who had breastfedor =12 months, vs. 0-1 months, during that interval). Our results suggest that thyroid stimulation during both pregnancy and lactation may result in a transient increase in risk of papillary thyroid cancer.

The authors determined the GSTM1 genotype of 190 Caucasian women with invasive squamous cell cervical cancer in western Washington State between January 1986 and June 1994 and of 206 controls. It was found that 53% of cases and 57% of controls had the GSTM1 null genotype. The age-adjusted odds ratio associated with GSTM1 null genotype was 0.8 (95% confidence interval 0.6, 1.3). The lack of association between cervical cancer risk and GSTM1 genotype was true both for smokers and nonsmokers of cigarettes and for heavy and light smokers. These data suggest that women with the GSTM1 null genotype are not at increased risk of cervical cancer. Am J Epidemiol 1999;150:568-72.

Glutathione S-transferases (GSTs) facilitate the excretion of a variety of potential carcinogens. Some 50-60% of Caucasians are homozygous for the null allele of GSTM1, a gene responsible for the presence of one of these enzymes. The authors examined whether women with the GSTM1 null genotype are at altered risk of vulvar cancer. They obtained peripheral blood specimens from 18- to 79-year-old residents of King, Pierce, and Snohomish counties of western Washington who were diagnosed with vulvar cancer between April 1991 and June 1994. Blood specimens were also obtained from controls identified via random digit telephone dialing of western Washington households. The authors determined the GSTM1 genotype of 137 cases (120 in situ and 17 invasive cases) and 248 controls. The frequency of the GSTM1 null genotype was 46.7% among cases and 57.3% among controls. The age-adjusted odds ratio associated with the GSTM1 null genotype was 0.7 (95% confidence interval: 0.4, 1.0). Among current smokers of cigarettes, the age-adjusted odds ratio associated with the GSTM1 null genotype was 0.5 (95% confidence interval: 0.2, 0.9), differing little between heavy and light smokers. Our data suggest that women with the GSTM1 null genotype are not at increased risk of vulvar cancer. Am J Epidemiol 1999;150:437-42.

Inconsistent results have been reported from studies evaluating the association of maternal smoking with birth of a Down syndrome child. Control of known risk factors, particularly maternal age, has also varied across studies. By using a population-based case-control design (775 Down syndrome cases and 7,750 normal controls) and Washington State birth record data for 1984-1994, the authors examined this hypothesized association and found a crude odds ratio of 0.80 (95% confidence interval 0.65-0.98). Controlling for broad categories of maternal age (35 years,or =35 years), as described in prior studies, resulted in a negative association (odds ratio = 0.87, 95% confidence interval 0.71-1.07). However, controlling for exact year of maternal age in conjunction with race and parity resulted in no association (odds ratio = 1.00, 95% confidence interval 0.82-1.24). In this study, the prevalence of Down syndrome births increased with increasing maternal age, whereas among controls the reported prevalence of smoking during pregnancy decreased with increasing maternal age. There is a substantial potential for residual confounding by maternal age in studies of maternal smoking and Down syndrome. After adequately controlling for maternal age in this study, the authors found no clear relation between maternal smoking and the risk of Down syndrome.

This study was conducted to assess the relation between body size and risk of breast cancer among young women. A case-control study was conducted among women aged 21-45 years living in three counties in Washington State. Cases were women born after 1944 with invasive or in situ breast cancer that was diagnosed between January 1, 1983, and April 30, 1990. Controls were selected using random digit dialing and were frequency-matched to cases on the basis of age and county of residence. Interviews took place between 1986 and 1992. Body size was evaluated using indices from several different time periods. After adjustment for confounders, a decreased risk of breast cancer was found for women in the highest quintile of body mass index (weight (kg)/height (m)2) as compared with the lowest quintile (for maximum lifetime body mass index, odds ratio = 0.69, 95% confidence interval (CI) 0.51-0.94). Age modified the relation between body size and risk of breast cancer. The odds ratio for women in the highest quintile of maximum body mass index who were aged 21-35 years was 0.29 (95% CI 0.16-0.55), as compared with an odds ratio of 1.5 for women aged 36-45 years (95% CI 0.9-2.5) (p for interaction = 0.003). This study supports prior research showing a decreased risk of breast cancer associated with increased body size among premenopausal or young women. More detailed analysis in this study found a strong effect that was limited to the youngest age group (< or = 35 years).

Some studies (but not all) have suggested that there may be an increase in the risk of breast cancer associated with a prior induced abortion. The risk, if present, may vary according to the duration of the pregnancy in which the abortion occurred, or to a woman's age or parity at that time. The authors conducted a case-control study of breast cancer in white women under age 45 years to address the question of breast cancer risk in relation to induced abortion, with the intention of identifying subgroups of women who might be at particularly increased risk. White women who were diagnosed with breast cancer (n = 1,302) from May 1, 1990, through December 31, 1992, in three geographic regions of the United States (Atlanta, Georgia; Seattle/Puget Sound, Washington; and five counties in central New Jersey) were interviewed about their reproductive histories, including the occurrence of induced abortion. Similar information was obtained from control women identified through random digit dialing. Logistic regression analysis was used to estimate the relative risk of breast cancer associated with a history of induced abortion, controlling for the potentially confounding influence of other breast cancer risk factors. Among women who had been pregnant at least once, the risk of breast cancer in those with a prior induced abortion was 20% higher than that in women with no history of abortion (95% confidence interval 1.0-1.5). This small increase in risk varied little according to number of abortions or a woman's current age. The association was present primarily among nulliparous women whose abortions occurred prior to 9 weeks' gestation (estimated relative risk = 2.0, 95% confidence interval 1.2-3.3). There was no excess risk of breast cancer associated with induced abortion among parous women. These data support the hypothesis that there may be a small increase in the risk of breast cancer related to a history of induced abortion among young women of reproductive age. However, the data from this study and others do not permit a causal interpretation at this time; neither do the collective results of the studies suggest that there is a subgroup of women in whom the relative risk associated with induced abortion is unusually high.Previous research has suggested that a history of induced abortion may be associated with an increased risk of breast cancer, but the risk may vary according to factors such as age, parity, or the timing of the abortion. To investigate these relationships further, a case control study was conducted among 1302 White women 20-44 years of age from three geographic areas in the US: Atlanta, Georgia; Seattle/Puget Sound, Washington; and five counties in central New Jersey). The women at these sites, who had been diagnosed with either in situ or invasive breast cancer between May 1, 1990, and December 31, 1992, met the demographic criteria and were available for interview; 1180 White age-matched controls were identified through random digit dialing. Women with breast cancer were more likely than controls to have a lower body mass and family history of breast cancer, to be nulliparous or have had fewer live births, to have had a late age at first birth or an early age at menarche, and to have used oral contraceptives for an extended time period. Among women who had ever been pregnant, the breast cancer risk in those with one or more induced abortion was 1.2 (95% confidence interval [CI], 1.0-1.5) relative to women with no abortion history. The breast cancer risk was greatest (2.0; 95% CI, 1.2-3.3) among nulliparous women whose abortion or abortions occurred prior to 8 weeks' gestation. This risk was slightly higher when the abortion was performed before 20 years of age (1.5) or after 29 years of age (1.4). The data from this study do not permit a causal interpretation at this time, nor do they identify any particular subgroup of women with induced abortion histories at enhanced risk of breast cancer.

A self-administered questionnaire was completed by 1,018 women diagnosed with breast cancer during 1988-1989 identified through the British Columbia Cancer Registry and by 1,025 controls selected at random from the Provincial Voters List. Parous premenopausal women who had never nursed (odds ratio (OR) = 1.3, 95% confidence interval (CI) 0.9-2.0) or who had lactated for 1 month or less (OR = 1.8, 95% CI 1.3-2.5) had an increased risk of breast cancer adjusted for age and parity, compared with women who had breast-fed 2 months or longer. The risk was particularly elevated (OR = 3.0, 95% CI 1.6-5.4) among women who reported having tried to nurse, but who were unsuccessful. Among women who nursed for at least 2 months, there was an indication of decreasing risk with increasing duration of nursing. Among postmenopausal parous women, no relation between lactation history and breast cancer risk was evident.

The association between cigarette smoking and cervical cancer has been demonstrated in numerous prior studies. As part of population-based case-control studies of cancers of the vulva, vagina, cervix, anus, and penis in relation to infection with human papillomavirus, conducted in western Washington State and the province of British Columbia from the mid 1980s until the present time, the authors have collected detailed information on smoking history. The proportion of subjects who were current smokers of cigarettes ranged from slightly over 40% among incident cases of vaginal and cervical cancer to 60% among cases of vulvar and anal cancer. In contrast, only about 25% of controls were current smokers. The adjusted odds ratios (OR) associated with current smoking were substantially elevated (OR = 1.9-14.6) for all cancer sites except cancer of the vagina (OR = 1.3). The risks tended to increase in proportion to the number of cigarettes smoked. For most cancer sites, the odds ratios associated with former smoking were substantially less than those associated with current smoking and diminished with increasing time since cessation of smoking. The authors' data and those of other investigators suggest that cigarette smoking plays a role in the etiology of anogenital cancers and that smoking has a late-stage or promotional effect.

To determine independent effects of maternal smoking and infant low birth weight (less than 2,500 g) on risk of sudden infant death syndrome (SIDS) among different ethnic groups, the authors conducted a population-based case-control study based on the 1984-1989 Washington State birth record data. Two control groups were selected for 916 SIDS cases. The first one comprised 3,704 randomly selected controls, matched to cases by birth year, to describe the characteristics of the study population. In the second control group (n = 6,186), minorities were oversampled, by matching to cases on maternal race/ethnicity and birth year, to increase the power of analysis within each ethnic group. All subjects were classified into five groups on the basis of maternal race/ethnicity: white, black, American Indian, Asian, and Hispanic. After controlling for confounders, the authors found that maternal smoking was independently associated with SIDS among white (odds ratio (OR) = 2.2, 95% confidence interval (CI) 1.8-2.6), blacks (OR = 3.1, 95% CI 1.7-5.9), Asians (OR = 2.7, 95% CI 1.1-6.6, and Hispanics (OR = 5.5, 95% CI 1.4-22.0), but had little relation among American Indians (OR = 1.4, 95% Cl 0.9-2.4). Infant low birth weight was independently related to SIDS among whites (OR = 2.5, 95% Cl 1.8-3.4) and American Indians (OR = 5.5, 95% Cl 2.8-11.2) and to a lesser extent among blacks (OR = 1.9, 95% Cl 0.8-4.1), but not among Asians (OR = 1.1, 95% Cl 0.2-5.2) or Hispanics (OR = 1.2, 95% Cl 0.1-11.5). The misclassification that may occur because of the application of the same definition of low birth weight to all ethnic groups may be the main reason for the weaker association between infant low birth weight and SIDS among blacks and the absence of an association among Asians and Hispanics. Defining low birth weight as below population mean minus 1.96 standard deviations may provide better insight into the relation between low birth weight and SIDS. Understanding the reasons for the lack of a strong association between maternal smoking during pregnancy and SIDS among American Indians may enhance our knowledge of the etiology and pathogenesis of SIDS.

In a case-control study of acute nonlymphocytic leukemia in western Washington state, cigarette smokers experienced a twofold increased risk of acute nonlymphocytic leukemia compared with nonsmokers. Risk increased significantly with the number of pack-years smoked (p = 0.0008) and decreased with the number of years stopped smoking (p = 0.15). The increased risk in smokers appeared to be limited to those who inhaled into the chest. The authors suggest that further studies of this potential relation are warranted, especially since the prevalence of cigarette smoking is relatively high in our society.

This case-control study was designed to investigate the relation between reproductive history and occurrence of hip and forearm fractures in postmenopausal women. Three hundred and fifty-five King County, Washington women who sustained a fracture between 1976 and 1980 and 562 control women were interviewed regarding their reproductive history and other factors. After controlling for the confounding effects of age, obesity, and use of noncontraceptive estrogens, the authors found that women who gave birth four or more times had a risk of fracture similar to women who had not given birth (odds ratio = 1.2, 95% confidence interval = 0.7-2.2), and women who breastfed for more than two years had a risk of fracture similar to women who had never breastfed (odds ratio = 0.8, 95% confidence interval = 0.4-1.5).

Female residents of western Washington state aged 18-80 years in whom thyroid cancer was diagnosed between January 1974 and December 1979 were interviewed concerning their reproductive histories and their prior use of exogenous estrogens. Their responses were compared with those of a sample of women from the same population, individually matched to cases on telephone prefix. Use of each of several estrogen-containing preparations was associated with a small increased risk of thyroid cancer; parous women who had ever used a lactation suppressant had 1.7 times the risk of parous nonusers (95% confidence interval, 1.1-2.8); ever users of oral contraceptives had 1.6 times the risk of never users (95% confidence interval, 0.98-2.5); and ever users of postmenopausal estrogens had 1.4 times the risk of never users (95% confidence interval, 0.89-2.3). Among the low risk group of women, i.e., those who had never undergone radiation therapy and who had never had a goiter, a history of one or more pregnancies was also associated with a small increase in the risk of thyroid cancer (relative risk = 1.8, 95% confidence interval, 1.1-3.1). However, no increase in risk with increasing duration of use of oral contraceptives or menopausal estrogens or with increasing number of pregnancies was noted. While pregnancy and use of exogenous estrogens have an impact on the production of thyroid-stimulating hormone, their effect on the incidence of thyroid carcinoma, if present at all, appears to be small.

The relationship of acute urinary tract infection occurring during pregnancy and the incidence of adverse fetal outcomes was examined with the use of birth certificate data from Washington state for the years 1980 and 1981. The two-year fetal mortality rate among urinary tract infection-associated pregnancies was found to be 17.9 per thousand livebirths, 2.4 times the rate for the state as a whole. Low birth weight newborns and small for gestational age newborns were, respectively, 2.04 and 1.57 times more frequent in urinary tract infection-associated pregnancies compared to controls. These estimates of risk were unchanged when adjusted for maternal age, race, and past obstetric history. The risk of prematurity in women with urinary tract infection and no history of fetal loss was 2.4 times the control risk. Sepsis was rare in both the urinary tract infection and control groups. These results add to the evidence implicating maternal urinary tract infection as a threat to fetal well-being.

As part of a population-based case-control study carried out in four areas of the United States, 427 cases of chronic lymphocytic leukemia diagnosed between July 1, 1977, and December 31, 1981, and 1,683 controls were interviewed regarding their history of chemical exposure. Respondents were asked if they had ever been "highly exposed" at home, at work, or elsewhere to one or more of a list of chemicals or to any other such substances not on the list. These chemicals were categorized into 20 exposure groups based on chemical composition. Odds ratios were calculated adjusting for age, sex, race, education level, and geographic location by means of unconditional logistic regression. Increased risks were associated with reported past exposure to acid-containing chemicals, "other caustic substances," aliphatic hydrocarbons, and chlorinated hydrocarbons. Because of the large number of exposures investigated in this study, and because of the relatively imprecise means the authors had to assess exposure, further research is needed to verify these findings.

Adult white male residents of 13 counties of western Washington State in whom germ cell testicular cancer was diagnosed between 1977 and 1983 (n = 333) were interviewed by telephone regarding their history of cryptorchism and its treatment. The same interview was given to a sample of 675 men selected from the population of these counties by dialing telephone numbers at random. Men who reported a history of cryptorchism were 5.9 times (95 per cent confidence interval 3.4-10.2) more likely than men without such a history to develop testicular cancer. Compared with noncryptorchid men, those with unilateral cryptorchism were at greater risk of developing a tumor on the side of nondescent (relative risk = 8.0) than on the opposite side (relative risk = 1.6). The size of the increased risk tended to be smaller among cryptorchid men who had undergone orchiopexy by age 10 than for other cryptorchid men, but the influence of orchiopexy in early childhood could not be evaluated in this population. These observations offer support for the hypothesis that one or more local factors (e.g., temperature elevation) account for the major part of the increased risk of germ cell testicular tumors in cryptorchid men.

To evaluate whether potent or prolonged stimulation of the immune system increases the risk of multiple myeloma, the authors compared 698 myeloma cases which occurred between July 1, 1977 and June 30, 1981 in four geographic areas of the United States with 1,683 demographically similar controls from the same areas. Cases and controls were interviewed about past exposures which may have involved antigenic challenge. Although few positive associations emerged, those most consistent with the immune stimulation hypothesis were modest associations between myeloma and a history of rheumatic fever (relative risk (RR) = 1.74, 95% confidence interval (CI) = 1.09-2.77) and between myeloma and urinary tract infection (RR = 1.30, 95% CI = 1.00-1.69, when self-respondent cases were compared with controls). Little association was found between the risk of myeloma and the number of past viral illnesses, number of bacterial illnesses, or number of allergy desensitization injections. Myeloma risk was found to be inversely related to the number of diseases against which a subject reported having been immunized, perhaps because of differences in socioeconomic status between cases and controls. These findings provide little support for the immune system stimulation hypothesis of myeloma etiology, but because of the limitations of interview techniques for assessing antigen exposure, further studies using laboratory methods may be warranted.

Adult male residents of 13 counties of western Washington state in whom testicular cancer had been diagnosed during 1977-1983 (n = 333) were interviewed over the telephone regarding their history of genital tract conditions, including vasectomy. For comparison, the same interview was given to a sample of 729 men selected from the population of these counties by dialing telephone numbers at random. A higher proportion of cases than controls reported having had a vasectomy (relative risk = 1.5, 95 per cent confidence interval = 1.0-2.2). However, the association was restricted entirely to Catholic men. Whereas a history of vasectomy was reported with approximately equal frequency by Catholic and non-Catholic cases, only 6.3 per cent of Catholic controls reported such a history in contrast to 19.7 per cent of other controls. While the authors cannot rule out the possibility that there is a true difference of the effect of vasectomy on the incidence of testicular cancer as a function of religion, it seems more plausible that selective underreporting by Catholic controls has produced a spurious relation.228 adult men resident of 13 western counties in Washington state, who had testicular cancer diagnosed between 1977-1983, were interviewed by telephone regarding vasectomy and other genitourinary history. 513 controls were selected at random from households chosen from the telephone directory by the Waksberg method. Only married men, aged 20-69, without history of cryptorchidism were included. Cancer patients and controls did not differ significantly in education, religion, smoking, coffee use or state residence. There was a small excess risk among men who reported history of inguinal hernia and testicular injury, and among those diagnosed with varicocele. Vasectomy was reported by 22.1% of the ever-married cases and 17.2% of controls. The relative risk of testicular cancer was the same for all groups except Catholics, whose risk was 8.7 times that of Catholic controls. There are no similar associations in the literature, either between testicular cancer and vasectomy, or between religion and under-reporting of contraceptive use, to explain this apparent under-reporting of vasectomy by Catholic controls.

The possibility that intrascrotal temperature is involved in the etiology of testicular cancer was investigated through a population-based case-control study conducted in western Washington State. A total of 323 men with germ cell tumors of the testis diagnosed between 1977 and 1984 and 658 randomly selected controls were interviewed with regard to type of shorts worn, use of long underwear, heat-resistant clothing, and hot tubs or saunas, and a history of varicocele. No association was found between testicular cancer and having worn tight-fitting underwear (jockey shorts) or having worn heat-resistant clothing on the job. A 50% increase in risk was observed for men who typically wore long underwear for more than three months out of the year, but this association was within the limits of chance given no true association. A smaller fraction of the men with cancer than controls had used a hot tub or sauna in the preceding five years, but the size of the case-control difference was not related to the frequency of use. The relative risk associated with a history of physician-diagnosed varicocele was 1.8 (95% confidence interval 0.9-3.4). There was little difference between men with seminomas and those with other germ cell tumors regarding any of the above exposures. Taken in aggregate, these results provide little or no support for the hypothesis that intermittent intrascrotal temperature elevation, to the degree encountered by US men today, plays a role in the etiology of germ cell testicular cancer. The possibility that risk increases after a continuous temperature increase, such as might result from a varicocele, needs to be evaluated further.

A nested case-control study was conducted to assess the relation between serum levels of selenium and retinol and the subsequent risk of cancer. During the years 1972-1984, in northwest Washington State, 156 cases of cancer were identified among members of two employee cohorts from whom specimens had been previously obtained and stored. Two hundred eighty-seven controls were selected from these cohorts and matched to cases on the basis of employer, age, sex, race, and date of blood draw. Selenium and retinol levels were measured by neutron activation and high pressure liquid chromatography, respectively. Information on known cancer risk factors was collected by telephone interviews of subjects and next of kin. Levels of selenium and retinol were unassociated with the incidence of cancer of all sites combined, both overall and within subgroups defined by age, sex, levels of the other micronutrient, time between blood draw and diagnosis, smoking status, and family history of cancer. These findings suggest that neither serum levels of selenium nor those of retinol have an appreciable effect on the risk of cancer.

Data from a population-based case-control study conducted in Washington State and Utah were used to assess whether obesity is associated with an altered risk of epithelial ovarian cancer. Quetelet index, defined as weight (kg) at age 30 years divided by height (m) squared, was calculated for each woman, and the values for all subjects were divided into five categories of approximately equal size. Compared with women in the lowest category, women in the highest category had an odds ratio of 1.7 (95 per cent CI 1.1-2.7). Risks for women in the three intermediate Quetelet index categories also exceeded the risk for women in the lowest group, but to a much smaller degree. Among women with serous tumors, those in the highest Quetelet index category were at a greater than twofold excess risk (OR = 2.2, 95 per cent CI 1.1-4.2), but the risk was not increased in the intermediate categories. For endometrioid tumors, risk increased consistently with increasing Quetelet index, and the odds ratio in the highest category was 4.7 (95 per cent CI 1.0-22.7). For both serous and endometrioid tumors, the excess risk was largely confined to premenopausal women. The results of this analysis suggest that for at least some types of ovarian tumor, obesity may warrant further attention as a possible etiologic factor.

Women resident in six counties of Washington and Utah for whom diagnoses of epithelial ovarian cancer were made during 1975-1979 were interviewed concerning their menstrual, reproductive and medical histories. For comparison, interviews were also obtained from a random sample of women living in the same countries. Logistic regression methods were used, and histories of childbearing, miscarriages, lactation, and (in Washington) usage of oral contraceptives were found to be associated with decreased risk of ovarian cancer; the estimated relative risks were, respectively, 0.88 per pregnancy (i.e., 0.88(2) for two pregnancies, etc.) (p = 0.016), 0.82 per miscarriage (p = 0.049), 0.79 per year of lactation (p = 0.034), and 0.89 per year of oral contraception (p = 0.009). In addition, it was observed that the magnitudes of the diminished risks from these exposures substantially exceeded those which would have been expected solely on the basis of their inhibition of ovulation (X2(5) = 21.5, p = 0.0006). On the other hand, the lack of association found between the occurrence of ovarian cancer and either total dose or total time of exposure to noncontraceptive estrogens, or with a history of usage of thyroid medications, suggests that periods of reduced pituitary gonadotrophin secretion fail to reduce risk of ovarian cancer. Thus, although pregnancy, lactation and oral contraception appear to offer some protection against the development of epithelial ovarian cancer, the reasons remain obscure.The association between factors of reproductive life and the occurrence of epithelial ovarian cancer were examined and anovulation and reduced gonadotrophin secretion were considered as mechanisms through which such factors might play their protective roles. All women with newly diagnosed epithelial ovarian cancer and who were resident in 6 counties of Washington and Utah during 1976 through 1979 were interviewed concerning their menstrual, reproductive, and medical histories. For comparison, interviews were also obtained from a random sample of women living in the same counties. To consider adequately the simultaneous effects of multiple relevant and possibly confounding variables, linear logistic regression techniques were used to analyze the data. Women with cancer reported fewer full-term pregnancies, fewer miscarriages, and less total time breastfeeding than controls. Cases in the Washington counties reported fewer exposures to combined oral contraceptive (OC) preparations. The difference between cases and controls was not apparent in the Utah data, possibly because of the low frequency of OC use among Utah residence and the small size of Utah samples. Obesity, defined as more than 20% excess weight at age 30 over the upper limit of ideal for a woman of a given height and medium frame was reported slightly more often by the cases than by the controls. Results obtained for reproduction variables appeared largely consistent with those of previous studies, in that factors associated with suppression of ovulation were generally protective. Histories of childbearing, miscarriages, lactation, and (in Washington) OC use were found to be associated with decreased risk of ovarian cancer. The estimated relative risks were, respectively, 0.88/pregnancy, 0.82/miscarriage, 0.79/year of lactation, and 0.89/year of OC. It was observed that the magnitudes off the diminished risks from these exposures substantially exceeded those which would have been expected solely on the basis of their inhibition of ovulation. The lack of association found between the occurrence of ovarian cancer and either total dose or total time of exposure to noncontraceptive estrogens, or with a history of usage of thyroid medications, suggests that periods of reduced pituitary gonadotrophin secretion fails to reduce risk of ovarian cancer. Pregnancy, lactation, and OC use appear to offer some protection against the development of epithelial ovarian cancer, yet the reasons remain obscure.

The authors undertook a case-control study of 113 cases of neonatal sepsis and 347 randomly selected controls. All cases and controls were selected from the 1980 and 1981 Washington State birth certificates. The increased risk for males (odds ratio (OR) = 1.75, p = 0.012) and the large risk associated with low birth weight (OR = 99.1, p less than 0.001 if less than 1,500 g and OR = 5.17, p less than 0.001 if 1,500-2,500 g) are consistent with past studies. The relationship of maternal age (OR = 2.00, p = 0.01 if less than or equal to 20 years and OR = 1.74, p = 0.05 if greater than 30) parallels the overall risk of many pregnancy-related complications in these age groups. Interpretation of the elevated risk associated with amniocentesis is hampered by small numbers but is provocative. The strong association of an Apgar score of 6 or less at five minutes (OR = 36.25, p less than 0.001) with neonatal sepsis suggests the possibility of routine sepsis evaluation in such neonates born in areas with high incidence rates of early neonatal sepsis. We found no previous reports associating either abruptio placentae (OR = 12.70, p = 0.028) or preeclampsia (OR = 6.43, p = 0.017) with neonatal sepsis.