The effect of recall bias on the association of calorie-providing nutrients and breast cancer.

This nested case-control study conducted within a large dietary cohort study examined whether recall bias could explain the inconsistent results obtained in case-control and cohort studies of the association between dietary fat and breast cancer. Cases were defined as women diagnosed with breast cancer between 1982 and 1987 who had completed a self-administered food frequency questionnaire on enrollment in the cohort study between 1982 and 1985. They were matched to two controls each, and the study subjects were asked in 1988 to complete a second diet questionnaire addressing their diets at the time of enrollment. The mean nutrient intakes for 325 cases who completed the first diet questionnaire six months or more before breast cancer diagnosis and their 628 matched controls were very similar for the prospectively and retrospectively collected diet questionnaires. There was little difference in the magnitude of the odds ratios estimated from the two questionnaires for the association between breast cancer risk and these nutrients. These data do not provide evidence for recall bias in retrospectively collected nutrient data from breast cancer cases compared with their matched controls.     

The validity of different control groups in a case-control study. Oral contraceptive use and breast cancer in young women

The possible selection bias due to the use of different series of controls in case-control studies has been analysed in the Norwegian part of a joint national case-control study in Sweden and Norway of oral contraceptive (OC) use and breast cancer in young women. A total of 105 cases diagnosed 1984-1985 were interviewed in addition to 210 population controls matched for age, 104 neighbourhood controls matched for age and postal address, and 96 hospitals controls matched for age. The relative risk was for each of the sets of controls 1.0 for 8 years and more of total use of OCs (adjusted for age at menarche, parity, age at first full term pregnancy, history of breast cancer in mother, history of benign breast biopsy). An additional questionnaire was sent to those who did not wish to participate (non-attenders). A slight selection bias was found for the population controls with more nulliparous never-users of OCs among the non-attenders than among interviewed controls.     

Case-control differences in the reliability of reporting a history of induced abortion

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.     

Risk of ectopic pregnancy and previous induced abortion.

OBJECTIVES: This study investigated the role of prior history of induced abortion in subsequent ectopic pregnancies. METHODS: Data from two French case-control studies were used to examine the effect of induced abortion on ectopic pregnancy risk. Case patients (n = 570) were women admitted for ectopic pregnancy during the study period; controls (n = 1385) were women who delivered in the same center. RESULTS: The analysis among women with no previous ectopic pregnancy showed that, after control for the main ectopic pregnancy risk factors, prior induced abortion was associated with an increased risk of ectopic pregnancy (odds ratio [OR] = 1.5, 95% confidence interval [CI] = 1.0, 2.0); there was a significant trend between number of previous induced abortions and ectopic pregnancy risk (ORs = 1.4 for 1 previous induced abortion and 1.9 for 2 or more). CONCLUSIONS: This study suggests that induced abortion may be a risk factor for ectopic pregnancy for women with no previous ectopic pregnancy, particularly in the case of women who have had several induced abortions.     

Pre- and postmenopausal breast cancer

It has been postulated that breast cancer is not a single disease, and that the risk factors occurring in premenopausal women are different from those occurring in postmenopausal women. This case-control study of pre- and postmenopausal breast cancer was designed to investigate a variety of variables including age of menarche, parity, age at first full-term birth, breast feeding patterns, age at menopause, and history of oral contraceptive use. The study compared 60 breast cancer cases and 125 matched controls from the same breast cancer screening population. Cases and controls were matched for race, age, place of birth, marital status, education, and annual family income. More cases than controls used oral contraceptives; the differences were statistically significant. The risk ratio estimates were 2.9 (confidence intervals 1.19-7.15). The mean duration of oral contraceptive use was more than two times longer among premenopausal breast cancer cases than controls and five times longer among postmenopausal cases than the controls. The breast cancer cases had significantly more relatives with a history of other cancers than the control group (risk ratio estimate 2.3, P < 0.03). No association was found between height and breast cancer; however, cases were found to be significantly heavier than controls. No differences were found in reproductive variables between breast cancer patients and their matched controls. In addition, a subdivision of breast cancer cases into pre- and postmenopausal groups did not reveal any clear differences in reproductive variables that would support the hypothesis that different specific risk factors are operating during pre- and postmenopausal periods.     

Diet and the risk of breast cancer in a case-control study: does the threat of disease have an influence on recall bias?

It has been suggested that recall bias may explain the discrepant results between case-control and cohort studies on diet and the risk of breast cancer. Two control groups were used for this case-control study of 25 to 75-year-old breast cancer cases (n = 310). The first group consisted of population controls drawn from the Finnish National Population Register (n = 454). The second group consisted of women who were referred to the same examinations as were the cases because of clinical suspicion of breast disease but who were later diagnosed as healthy (referral controls; n = 506). Because the diagnosis was unknown at the time of interview, it was possible to assess by comparing the two control groups whether the self-reporting of diet changed under the threat of disease. Dietary habits were examined using a validated, self-administered food-frequency questionnaire. Premenopausal women misreported their consumption of liquid milk products, tea, and sugar. Reporting bias was also associated with the intake of fat and vitamins. Postmenopausal women misreported consumption of milk products. When recall bias was taken into consideration, milk was associated with increased risk of premenopausal breast cancer, whereas high consumption of poultry or high intake of monounsaturated fatty acids, n-3 fatty acids, n-6 fatty acids, and vitamin E were related to lower risk. The study suggested that oil, milk, cheese, coffee and beta-carotene may act as protective factors in postmenopausal women, whereas butter and cream may be risk factors for breast cancer. In summary, it is possible that some food items may be overreported or underreported under the threat of disease in health-conscious population. However, most of the results in this study were not modified by recall bias.     

A case-control evaluation of the effect of breast cancer screening in the United Kingdom trial of early detection of breast cancer.

OBJECTIVE--The aim was to assess the extent to which selection bias affects a case-control study of breast cancer screening in which attenders and non-attenders for screening are compared. DESIGN--There were two retrospective case-control studies, one estimating the risk of death from breast cancer in women in the screening district relative to those in the comparison district (study A), the second estimating the relative risk for women who had ever been screened compared with women who had never been screened in the screening district alone (study B). For cases and controls in study B, the women's screening history was summarised for the time period from date of entry to diagnosis of the case, or the equivalent time from date of entry for the matched controls. For cases detected by screening, the screen at which cancer was detected was included in the screening history. SUBJECTS--Cases were deaths from breast cancer in women with disease diagnosed after entry to the trial, up to 31 December 1986 or a maximum of seven years from date of entry, in one of the screening districts (Guildford) and one of the comparison districts (Stoke) participating in the UK Trial of Early Detection of Breast Cancer: study A: 198 deaths in Guildford and Stoke; study B: 51 deaths in Guildford only. There were five age matched controls for each case, with length of follow up at least as great as the time from entry to death of the case. MAIN RESULTS--The estimate of the risk of death from breast cancer in the screening district relative to the comparison district from study A was 0.76, thus implying a reduction of 24% in the screening district, similar to that obtained from a cohort analysis of data from the two districts. In contrast, the relative risk in study B for ever v never screened women was 0.51, which, taking the 72% compliance into account, would result in a relative risk of 0.65 for the screening district if there were no selection bias. The risk of breast cancer mortality in the never screened relative to the comparison district was 1.13, despite the fact that incidence rates in the two populations were similar. This suggested that cancers in the never screened group had a particularly poor prognosis, contributing to selection bias. CONCLUSIONS--The possible existence of selection bias should lead to caution in interpretation of the results of case-control studies of the effect of breast cancer screening on mortality.     

Potential for bias in case-control studies of oral contraceptives and breast cancer

The discrepancies between the findings of the 6 large case-control studies to study the association between oral contraceptive (OC) use and breast cancer diagnosed in the 1980s may be due to chance or bias. The likelihood that chance played a role is suggested by the large numbers of subgroups examined in each study, the inconsistencies in the findings of different studies, and the wide confidence intervals around most of the relative risks. The most serious potential problem in case-control studies is that the procedures used to select cases and controls may produce groups that are not truly comparable. Recall bias is likely to contaminate information about the duration and type of part OC use. In addition, the more frequent examination of the breasts of women using OCs can produce surveillance bias. 6 procedures are recommended to minimize bias in future case-control studies of OC use and breast cancer: 1) whenever possible, cases and controls should be selected from an entire community; 2) if hospital controls need to be used, there should be explicit criteria for selecting them and the proportions of OC users in each diagnostic group should be presented; 3) women interviewed should not be aware of the study's hypotheses; 4) interviewers should be kept blind as to whether a subject is a case or control; 5) the possibility of recall bias should be investigated by comparing contraceptive histories from a sample of cases and controls with an independent source of information (preferably recorded before cancers were diagnosed); and 7) the interview should include questions about the frequency of breast examinations, so that any effects of more frequent surveillance of OC users can be controlled for.     

Induced abortion and the chromosomal characteristics of subsequent miscarriages (spontaneous abortions)

Data from a case-control study of miscarriages (spontaneous abortions) were used to test whether single and multiple induced abortions are associated with miscarriage in subsequent euploid (chromosomally normal) pregnancies. Cases and controls were identified in three New York City hospitals between April 1974 and November 1982. It was hypothesized that, if induced abortion increased the risk of subsequent miscarriage, an association would be observed with euploid but not aneuploid (chromosomally abnormal) miscarriage. The frequencies of single and multiple induced abortions among euploid cases and aneuploid cases were compared with those among controls. Among both private and public patients, the proportion of women reporting a single induced abortion was similar in euploid cases and in aneuploid cases compared to controls. Among public patients only, the proportions reporting multiple induced abortions were also similar in euploid cases and in aneuploid cases compared to controls (odds ratios = 0.9 and 1.0, respectively). In contrast, among private patients, the proportion reporting multiple induced abortions was raised among euploid cases (odds ratio = 2.2, 95% confidence interval = 1.3-3.7), although not among aneuploid cases. This association was strongest when the first induced abortion was carried out before 1973, at a young age, or with a procedure other than suction curettage. In public patients, the associations with miscarriage did not vary with characteristics of the first induced abortion, but multiple induced abortion histories when the first two induced abortions occurred before 1973 were in excess among euploid cases compared to controls. These results suggest that, for both private and public patients, neither single nor multiple induced abortions as now performed are likely to increase the risk of miscarriage in subsequent pregnancies. They also suggest a mechanical origin for some miscarriages.     

Oral contraceptive use and premenopausal breast cancer in Sweden and Norway: possible effects of different pattern of use

Different use of oral contraceptives (OCs) in relation to reproductive histories and age might explain the conflicting results found in studies from different countries of OC use and premenopausal breast cancer. A population-based case-control study in Sweden and Norway has been analysed separately for the two countries. The study consists of 317 Swedish and 105 Norwegian cases diagnosed 1984-1985 with 317 Swedish and 210 Norwegian controls. The results for each country separately are consistent with the joint analysis showing increased risk for breast cancer with increased duration of OC use with the exception of that for women with more than 15 years since first use. The relative risk of breast cancer for more than four years of use among Norwegian women was 0.6 (95% confidence internal (Cl), 0.2-1.8). In both countries there was a non-significant higher risk associated with increasing duration of use before first full-term pregnancy. Analyses for surveillance bias show no evidence of bias as explanation for the association between total duration of OC use and breast cancer.     

Associations between breast cancer, plasma triglycerides, and cholesterol

A case-control study investigating the association between plasma lipids and breast cancer was conducted among women aged 30-80 in Buffalo, NY. All eligible women from a large breast clinic and two area physicians' offices were requested to participate over a one-year period. Subjects completed a health questionnaire and donated a fasting blood sample prior to diagnostic breast biopsies. The 83 women found to have breast cancer (cases) had significantly higher plasma triglyceride values than did the 113 women found not to have breast cancer (controls). Lower plasma beta-carotene values were associated with breast cancer, but only in those women with elevated triglyceride or cholesterol. Plasma cholesterol values were lower in those breast cancer cases presenting with more advanced stages of cancer, suggesting that metabolic effects of clinical and preclinical breast cancer may lower cholesterol levels. Although the limitations of case-control studies are well-recognized, these data suggest an etiologic role for plasma triglycerides and beta-carotene or for related dietary factors.     

Infertility and breast cancer: a population-based case-control study

To investigate whether a history of infertility affects a woman's risk of developing breast cancer, the authors analyzed case-control data collected between 1980 and 1982 as part of the Cancer and Steroid Hormone Study. The 4,730 cases were women aged 20-54 years with a first diagnosis of breast cancer ascertained from eight population-based cancer registries; the 4,688 controls were women randomly selected from the general population of these same eight areas. After controlling for age, age at first birth, and parity, the odds ratio (OR) for breast cancer associated with infertility was 1.01 (95% confidence interval (CI) 0.89-1.15) among gravid women. Controlling for age, the odds ratio was 0.82 (95% CI 0.59-1.14) among nulligravid women. Women who reported that the reason for their infertility was a problem with their ovaries had a risk similar to that for women without a history of infertility (OR = 0.75, 95% CI 0.48-1.24). Women whose physicians reported that the reason for their infertility was anovulation or Stein-Leventhal syndrome also had risks similar to those for women without a history of infertility (OR = 1.26 (95% CI 0.67-2.34) and OR = 1.13 (95% CI 0.46-2.78), respectively). Menopausal status, age at menarche, history of spontaneous abortions, drinking or smoking behavior, use of exogenous hormones, or family history of breast cancer did not appreciably alter the observed odds ratios. If infertility has an effect on breast cancer that is independent of age at first birth, then the effect is small.     

Agreement between oral contraceptive users and prescribers: implications for case-control studies.

Case-control studies examining the effects of oral contraceptives (OC) are prone to misclassification bias due to errors in assessment of OC use. Concern about inaccurate exposure histories has increased since current studies require women to recall OC use over prolonged periods of time. In preparation for a case-control study of breast cancer and OC use, an investigation was carried out to assess agreement between women's lifetime histories of OC use (covering a period of up to 20 years) and prescribers' records. OC histories were obtained during personal interview with 218 women who had used OC at some point in their lives (127 breast cancer patients, 91 controls). Recall was aided by an album with color photographs of all OC marketed in the Netherlands from 1962 onwards (n = 65), and a calendar that covered the women's life span from date of birth to menopause. The participants were asked for the names of all physicians who prescribed OC for them. The rate of response from the prescribers was high (94%), but only half of the forms provided useful information. Patient-prescriber agreement on brand names (including dosage) was 70%. About half of the women agreed with their prescribers on starting dates to within less than a year's difference. Approximately the same percentage of agreement was found for stopping dates. Multiple linear regression indicated that agreement on brand names and dates of usage was lower for women of low socioeconomic status, for healthy women (as compared to breast cancer patients) and for periods of pill use that had to be recalled from the more distant past. Agreement on total duration of use was high enough to permit testing of a moderately strong duration-response relationship in a case-control study.     

Incomplete pregnancy is not associated with breast cancer risk: the California Teachers Study

BACKGROUND: Early studies of incomplete pregnancy and development of breast cancer suggested that induced abortion might increase risk. Several large prospective studies, which eliminate recall bias, did not detect associations, but this relationship continues to be debated. STUDY DESIGN: To further inform this important question, we examined invasive breast cancer as it relates to incomplete pregnancy, including total number of induced abortions, age at first induced abortion and total number of miscarriages among women participating in the ongoing California Teachers Study (CTS) cohort. Incomplete pregnancy was self-reported on the CTS baseline questionnaire in 1995-1996. Incident breast cancers were ascertained in 3324 women through 2004 via linkage with the California Cancer Registry. RESULTS: Using Cox multivariable regression, we found no statistically significant association between any measure of incomplete pregnancy and breast cancer risk among nulliparous or parous women. CONCLUSION: These results provide strong evidence that there is no relationship between incomplete pregnancy and breast cancer risk.     

Breast cancer risk factors among black women and white women: similarities and differences.

To further explore whether breast cancer risk factors are the same for black women and white women, the authors investigated several biologic, cultural, and social factors in a 1980-1982 case-control study of non-Hispanic black subjects (490 cases, 485 controls) and non-Hispanic white subjects (3,934 cases, 3,901 controls) aged 20-54 years. Logistic regression analyses indicated that blacks and whites shared four risk factors at a comparable magnitude (age at first full-term birth, parity, surgical menopause, and history of benign breast disease). For two observed risk factors, the magnitude (breast feeding) and pattern (family history of breast cancer) of the relation were different in blacks and whites. The relative risks of breast cancer among black women who had first-degree relatives (odds ratio (OR) = 1.61) and second-degree relatives (OR = 1.71) with breast cancer were similar, whereas the relative risk among white women who had first-degree relatives (OR = 2.16) was distinctly larger than for those who had second-degree relatives (OR = 1.44) with breast cancer. The relation of early age at menarche appeared negligible for blacks although significant for whites aged 12 and under (OR = 1.26). The results also indicated that natural menopause, oral contraceptive use, and cigarette smoking may have different, and more complex, relations to breast cancer among blacks and whites.     

Spontaneous abortions among women working in the pharmaceutical industry.

A register based study was conducted on the pregnancy outcome of female workers in eight Finnish pharmaceutical factories to determine whether they had a higher risk of spontaneous abortion than the general population or matched controls. Information about all female workers who had been employed in the factories during the years 1973 or 1975 (four factories) to 1980 was obtained from the employers. The workers' pregnancy data were collected from the nation wide hospital discharge register and polyclinic data of hospitals from 1973 to 1981. The total number of 1795 pregnancies included 1179 deliveries, 142 spontaneous abortions, and 474 induced abortions. The spontaneous abortion rate (the number of spontaneous abortions X 100, divided by the number of spontaneous abortions plus the number of births) during employment was 10.9% and before/after employment 10.6%. The rate for all the women in the corresponding central hospital districts was 11.3% [corrected] during the study period. A case-control study was also carried out in which the cases were 44 women who had a spontaneous abortion during employment in the pharmaceutical factory. Three age matched female pharmaceutical factory workers who had given birth to a child were chosen as controls for every case. The information about occupational exposures was collected from questionnaires completed by the occupational physician or nurse at the factory. The response rate was 93%. Exposure to chemicals was more common among the cases than among the controls. For methylene chloride, a solvent commonly used in the pharmaceutical industry, the increase in odds ratio of borderline significance (odds ratio 2.3, p = 0.06). In a logistic regression model (which included oestrogen exposure, solvent exposure frequency of the usage, and heavy lifting) the odds ratio was increased for oestrogens (odds ratio 4.2, p = 0.05) and for continuous heavy lifting (odds ratio 5.7, p = 0.02). The odds ratio for spontaneous abortions was greater among those exposed to four or more solvents (odds ratio 3.5, p=0.05) than among those exposed to one to three solvents (odds ration 0.8, p=0.74).     

Spontaneous abortions among women working in the pharmaceutical industry

A register based study was conducted on the pregnancy outcome of female workers in eight Finnish pharmaceutical factories to determine whether they had a higher risk of spontaneous abortion than the general population or matched controls. Information about all female workers who had been employed in the factories during the years 1973 or 1975 (four factories) to 1980 was obtained from the employers. The workers' pregnancy data were collected from the nation wide hospital discharge register and polyclinic data of hospitals from 1973 to 1981. The total number of 1795 pregnancies included 1179 deliveries, 142 spontaneous abortions, and 474 induced abortions. The spontaneous abortion rate (the number of spontaneous abortions X 100, divided by the number of spontaneous abortions plus the number of births) during employment was 10.9% and before/after employment 10.6%. The rate for all the women in the corresponding central hospital districts was 11.3% [corrected] during the study period. A case-control study was also carried out in which the cases were 44 women who had a spontaneous abortion during employment in the pharmaceutical factory. Three age matched female pharmaceutical factory workers who had given birth to a child were chosen as controls for every case. The information about occupational exposures was collected from questionnaires completed by the occupational physician or nurse at the factory. The response rate was 93%. Exposure to chemicals was more common among the cases than among the controls. For methylene chloride, a solvent commonly used in the pharmaceutical industry, the increase in odds ratio of borderline significance (odds ratio 2.3, p = 0.06). In a logistic regression model (which included oestrogen exposure, solvent exposure frequency of the usage, and heavy lifting) the odds ratio was increased for oestrogens (odds ratio 4.2, p = 0.05) and for continuous heavy lifting (odds ratio 5.7, p = 0.02). The odds ratio for spontaneous abortions was greater among those exposed to four or more solvents (odds ratio 3.5, p=0.05) than among those exposed to one to three solvents (odds ration 0.8, p=0.74).     

Cigarette smoking and breast cancer.

An epidemiological case-control study was conducted in New York State, with 1617 primary breast cancer patients and an equal number of controls, to examine the relationship between cigarette smoking and breast cancer. Results showed no overall association between ever smokers versus never smokers and breast cancer risk (odds ratio [OR] = 1.03, 95% confidence interval [CI]: 0.90-1.19), nor was there any dose response trend observed with increased levels of smoking. In addition, no association was found with risk and age started smoking, age stopped smoking, amount smoked or total years smoked. Controlling for previously identified risk factors for breast cancer in the analysis did not significantly alter these relationships. Previous studies have found a difference in menopausal age among smokers compared to nonsmokers. The mean menopausal age was only slightly lower in smokers than in never smokers for both cases and controls. Breast cancer risk was observed to be close to unity for premenopausal women (OR = 0.97, 95% CI: 0.74-1.34) and postmenopausal women (OR = 1.06, 95% CI: 0.91-1.26). A recent study suggested breast cancer risk was more strongly related to starting smoking at a young age among women who smoked at least 25 or more cigarettes per day in the most recent year of smoking. This hypothesis was not supported by these data.     

Participation rates in a case-control study: the impact of age, race, and race of interviewer

PURPOSE: Despite concerns about declining participation rates in epidemiologic studies in recent years, relatively few papers have discussed obstacles to recruiting study participants or strategies for optimizing response rates. This report describes factors associated with nonparticipation in a population-based, case-control study of breast cancer and discusses ways to overcome barriers to participation. METHODS: Contact and cooperation rates were calculated for participants in the Carolina Breast Cancer Study (CBCS), stratified by case status, age, race, and race of interviewer. Demographic and breast cancer risk factor characteristics of partial and full responders also were compared. RESULTS: Contact rates and cooperation rates varied by case/control status and demographic characteristics. Contact rates were lower among controls, younger women, and black women. Cooperation rates were lower among controls, older women, and black cases. Cooperation rates were higher among both black and nonblack women when participants and interviewers were concordant on race. CONCLUSIONS: Obstacles to recruitment seem to differ among race and age subgroups, suggesting that recruitment strategies may need to be tailored to potential participants based upon demographic characteristics. Strategies have been implemented to improve response rates in this and other epidemiologic studies; however, additional research and innovation in this area are needed.     

Preeclampsia, pregnancy-related hypertension, and breast cancer risk

Pregnancy conditions accompanied by high blood pressure, such as preeclampsia and pregnancy-related hypertension, have been associated with a lower risk of breast cancer in several epidemiologic studies. It is unknown whether length of gestation or multiple occurrence of these conditions alters the association with breast cancer. It is also unknown whether the inverse association between preeclampsia and breast cancer risk is modified by menopausal status at breast cancer diagnosis. Using data from a large, population-based case-control study of breast cancer conducted on Long Island, New York, during 1996-1997, the authors examined these questions among ever-parous women (1,310 cases and 1,385 controls) using multivariate logistic models. Preeclampsia was inversely associated with breast cancer (odds ratio = 0.7, 95% confidence interval: 0.5, 1.0); this association was even stronger among women who had multiple occurrences of preeclampsia (odds ratio = 0.3, 95% confidence interval: 0.1, 0.9). The risk reduction was more pronounced among postmenopausal women. Gestation length did not substantially alter the relation between preeclampsia and breast cancer risk. Pregnancy-related hypertension was also inversely associated with breast cancer risk, but the relations were not statistically significant after adjustment for preeclampsia. These data suggest that pregnancy conditions related to hypertension, particularly preeclampsia, play a role in reducing breast cancer risk. Possible biologic mechanisms underpinning these associations should be further explored.