Exploring associations between residential location and breast cancer incidence in a case-control study

Locating geographic hot spots of cancer may lead to new causal hypotheses and ultimately to new knowledge of cancer-causing factors. The Cape Cod region of Massachusetts has experienced elevated incidence of breast cancer compared with statewide averages. The origins of the excess remain largely unexplained, even after the Upper Cape Cod Cancer Incidence Study investigated numerous potential environmental exposures. Using case-control data from this study (258 cases and 686 controls), we developed an exploratory approach for measuring associations between residential location and breast cancer incidence, adjusting for individual-level risk factors. We measured crude and adjusted odds ratios over the study region using fixed-scale grids and a smoothing algorithm of overlapping circular units. Polycircular hot spot regions, derived from the peak values of the smoothed odds ratios, delineated geographic areas wherein residence was associated with 60% [odds ratio (OR), 1.6; 95% confidence interval (CI), 0.8-3.2] to 210% (OR, 3.1; 95% CI, 1.3-7.2) increased incidence relative to the remainder of the study population. The findings suggest several directions for further research, including the identification of potential environmental exposures that may be assessed in forthcoming case-control studies.     

Potential misinterpretation of the case-only study to assess gene-environment interaction.

Novel epidemiologic study designs are often required to assess gene-environment interaction. A design using only cases, without controls, is one of several approaches that have been proposed as more efficient alternatives to the typical random sampling of cases and controls. However, it has not been pointed out that a case-only analysis estimates a different interaction parameter than does a traditional case-control analysis: The latter typically estimates departure from multiplicative population odds or rate ratios, depending on the method of control selection, while the former estimates departure from multiplicative risk ratios if genotype and environmental exposure are not associated in the population. These parameters are approximately equal if the disease risk is small at all levels of the study variables. The authors quantify the impact of allowing for higher disease risk among gene carriers, a relevant situation when the gene under study is highly penetrant. Their findings show that the cross-product ratio computed from case-only data may be substantially smaller than the odds ratio computed from case-control data and may therefore underestimate either the population odds or the rate ratio. Thus, to avoid misinterpretation of interaction parameters estimated from case-only data, the definition of multiplicative interaction should be made explicit.     

Sample size evaluation for a multiply matched case-control study using the score test from a conditional logistic (discrete Cox PH) regression model

The conditional logistic regression model (Biometrics 1982; 38:661-672) provides a convenient method for the assessment of qualitative or quantitative covariate effects on risk in a study with matched sets, each containing a possibly different number of cases and controls. The conditional logistic likelihood is identical to the stratified Cox proportional hazards model likelihood, with an adjustment for ties (J. R. Stat. Soc. B 1972; 34:187-220). This likelihood also applies to a nested case-control study with multiply matched cases and controls, selected from those at risk at selected event times. Herein the distribution of the score test for the effect of a covariate in the model is used to derive simple equations to describe the power of the test to detect a coefficient theta (log odds ratio or log hazard ratio) or the number of cases (or matched sets) and controls required to provide a desired level of power. Additional expressions are derived for a quantitative covariate as a function of the difference in the assumed mean covariate values among cases and controls and for a qualitative covariate in terms of the difference in the probabilities of exposure for cases and controls. Examples are presented for a nested case-control study and a multiply matched case-control study.     

Differences in risk estimations from a hospital and a population-based case-control study

A hospital-based and population-based case-control study of cervical cancer (in situ and invasive) was conducted in urban Utah to determine if methods of respondent selection affect estimates of risk for variables thought to be associated with the disease. Population cases (N = 409) and cases from two large hospitals (N = 124) were identified through the Utah Cancer Registry. Population-based controls (N = 379) were identified through random-digit dialing; hospital-based controls (N = 150) with gynaecological disorders other than cancers and elective abortions were chosen from the same hospitals as the cases for the hospital study. Both control groups were frequency matched to cases by age. Approximately 79% of the identified cases and 85% of the selected controls completed interviews conducted in their homes. Most risk estimates were lower in the hospital-based study because of the more case-like attributes of this group. Stratified analysis for social class led to adjusted risk estimates which were lower than the unadjusted risk estimates for the population-based study, but not for the hospital-based study. The close social class matching in the hospital-based study seems to have led to concurrent overmatching on other risk factors since many of these are closely related to social class. Findings are discussed in terms of implications for case-control study design.     

An empirical study of the use of living versus deceased study subjects: associations with liver cancer in the selected cancers study.

PURPOSE: In case-control studies, the issue of whether living or deceased controls should be used for deceased cases has been controversial. METHODS: Using data from a study of cancer among men that selected both live (n = 1910) and deceased controls (n = 596) for cases of liver cancer (109 deceased, 59 living), we examined the effects of using information from proxy respondents (cases and controls) and from live cases and controls on associations between liver cancer and known risk factors. Cases diagnosed between 1984 and 1988 were selected from eight population-based cancer registries. Live controls were recruited by random digit dialing, deceased controls from death certificate files. Controls were matched to cases on geographic area, year-of-birth, and race. RESULTS: Adjusted odds ratios (OR) calculated for deceased cases and controls, when compared to odds ratios for live cases and controls, were attenuated towards the null value for history of hepatitis (4.7 vs. 14.9), blood transfusions (1.1 vs. 7.8), and cirrhosis (9.3 vs. 51.1). When all cases and living controls were used, odds ratios did not differ substantially from those for living cases and controls except for cirrhosis (OR = 154.2). For smoking, the odds ratios were similar in all analyses. Adjustment for type of interview (self, proxy) did not eliminate differences between results for living and deceased subjects; significant interactions were found between type of interview and hepatitis, cirrhosis, and blood transfusions. CONCLUSIONS: Selection of live controls for deceased cases is recommended to decrease misclassification in measures of exposure.     

Breast cancer in men: risk factors with hormonal implications.

Cases included in a population-based case-control study of breast cancer in men were recruited from 10 geographic areas of the United States from 1983 to 1986. Controls, matched to cases on age and geographic area, were selected by random digit dialing for men under age 65 years and from Health Care Financing Administration files for older men. Results are based on responses from 227 cases and 300 controls to questions asked in a standardized personal interview. An increased risk of breast cancer was most strongly associated with undescended testes and was also related to orchiectomy, orchitis, testicular injury, late puberty, and infertility; and a decreasing trend in risk was observed with an increasing number of children. Relative risk estimates were also elevated in relation to a history of high blood cholesterol, rapid weight gain, benign breast conditions, and possibly obesity. These findings suggest that breast cancer in men develops in response to androgen deficiency associated with testicular dysfunction and under conditions associated with excess estrogen. Risk was also found to be elevated in men with a history of amphetamine use, diabetes, and cigar smoking and reduced in men with prior head trauma.     

Escherichia coli O157:H7 outbreak associated with consumption of ground beef, June-July 2002

OBJECTIVE: A case-control and environmental study tested the hypothesis that purchasing and eating ground beef from a specific source was the cause of a cluster of cases of hemolytic uremic syndrome (HUS) and Escherichia coli (E. coli) O157:H7 gastroenteritis. METHODS: A case-control study comparing risk factors was conducted over the telephone on nine case-patients with 23 selected controls. An environmental investigation was conducted that consisted of reviewing beef handling practices at a specific local supermarket and obtaining ground beef samples from the store and two households with case-patients. RESULTS: The analysis of the case-control study showed that eight case-patients (89%) purchased ground beef at Grocery Chain A compared with four controls who did not develop illness (17%) (matched odds ratio=undefined; 95% confidence interval 2.8, infinity; p=0.006). The environmental investigation showed that Grocery Chain A received meat from Meatpacker A. Laboratory analysis of meat samples from Meatpacker A and Grocery Chain A and stool samples from some patients recovered an identical strain of E. coli O157:H7 according to pulse-field gel electrophoresis. CONCLUSIONS: Both the case-control and environmental studies showed that purchasing ground beef at Grocery Chain A, which received ground beef from Meatpacker A, was the major risk factor for illness in eight case-patients; the ninth case-patient was found to be unrelated to the outbreak. Furthermore, meat from Meatpacker A was associated with a nationwide outbreak of E. coli O157:H7 illness that resulted in the second largest recall of beef in U.S. history at the time.     

Epithelial ovarian cancer and coffee drinking

A recent report from a case-control study in Greece suggested that coffee consumption is related to an increase in the risk of ovarian cancer. This hypothesis was examined in a hospital-based case-control study in the US. Information on coffee drinking and other factors was obtained from 290 incident cases of epithelial ovarian cancer and compared with that of 580 controls with non-malignant conditions of acute onset and 476 controls with cancer of other sites. Adjustment was made for the potential confounding effects of several factors, including the major known risk factors for ovarian cancer. The estimated relative risk for drinking five or more cups of coffee per day, relative to none, was 1.1 (95% confidence interval, 0.6-2.0) using the controls with non-malignant conditions and 1.0 (0.5-1.8) using the controls with cancer. The estimates for drinking less than five cups per day were greater than 1.0, but this could have been due to chance. The consumption of decaffeinated coffee and tea also appeared to have no influence on risk. The evidence from this study suggests that, if coffee drinking increases the risk of ovarian cancer, the effect is small.     

Outbreak of a new Salmonella phage type in South West England: alternative epidemiological investigations are needed

An outbreak of Salmonella Typhimurium DT193a occurred in South West England in April-May 2003. Preliminary investigation suggested an association with the consumption of ham. A matched case-control study was implemented, together with an environmental investigation and a geographic information system (GIS) analysis. Thirty-seven cases and 38 controls were enrolled in the study. Matched analysis was based only on 23 cases, as the other cases did not nominate controls. Eighty per cent of cases and 51% of controls had eaten ham (odds ratio = 3.5, p = 0.03). Cases did not differ from controls in terms of distance from outlets providing ham. All environmental samples collected were negative. The epidemiological evidence indicated an association between this outbreak and consumption of ham, but the environmental investigation was inconclusive. Alternative methods should be considered in support of traditional epidemiological investigation.     

Selection of controls in case-control studies. I. Principles.

A synthesis of classical and recent thinking on the issues involved in selecting controls for case-control studies is presented in this and two companion papers (S. Wacholder et al. Am J Epidemiol 1992;135:1029-50). In this paper, a theoretical framework for selecting controls in case-control studies is developed. Three principles of comparability are described: 1) study base, that all comparisons be made within the study base; 2) deconfounding, that comparisons of the effects of the levels of exposure on disease risk not be distorted by the effects of other factors; and 3) comparable accuracy, that any errors in measurement of exposure be nondifferential between cases and controls. These principles, if adhered to in a study, can reduce selection, confounding, and information bias, respectively. The principles, however, are constrained by an additional efficiency principle regarding resources and time. Most problems and controversies in control selection reflect trade-offs among these four principles.     

Investigation of environmental and host-related risk factors for tuberculosis in Africa. I. Methodological aspects of a combined design

Host-related and environmental factors for tuberculosis have usually been investigated separately using different study designs. Joint investigation of the genetic, immunologic, and environmental factors at play in susceptibility to tuberculosis represents an innovative goal for obtaining a better understanding of the pathogenesis of the disease. In this paper, the authors describe methods being used to investigate these points in a West African study combining several designs. Patients with newly diagnosed smear-positive cases of tuberculosis are recruited. The effect of host-related factors is assessed by comparing each case with a healthy control from the case's household. The role of environmental factors is estimated by comparing cases with randomly selected community controls. The frequencies of candidate gene variants are compared between cases and community controls, and results are validated through family-based association studies. Members of the households of cases and community controls are being followed prospectively to determine the incidence of "secondary" tuberculosis and to evaluate the influence of geographic and genetic proximity to the index case. This type of design raises important methodological issues that may be useful to consider in studies investigating the natural history of infectious diseases and in attempts to disentangle the effects of environmental and genetic factors in response to infection.     

Multiple births and maternal risk of breast cancer

Data from the Cancer and Steroid Hormone Study, a large nationwide population-based case-control study conducted in the United States in 1980-1982, were analyzed to investigate whether pregnancies ending in a multiple birth affect the risk of subsequent breast cancer. The cases were 3,918 parous women who were aged 20-54 years and newly diagnosed with breast cancer; controls were 4,047 parous women selected randomly from the same geographic areas as the cases. Multiple births were reported by 118 cases and 161 controls. After adjustment for other reproductive variables, having a multiple last birth was found to be protective against breast cancer (odds ratio (OR) = 0.60, 95% confidence interval (CI) 0.43-0.85), whereas having a multiple birth prior to the last birth was not (OR = 1.11, 95% CI 0.79-1.57). To the authors' knowledge, this study is the first investigation to report such a protective effect, and thus the finding warrants replication. One mechanism that might account for the effect involves the increased output of alpha-fetoprotein by multiple fetal livers.     

Oral contraceptive use and liver cancer

The risk of liver cancer in relation to use of oral contraceptives was evaluated in a hospital-based case-control study conducted in five US cities from 1977 to 1985. Twelve new cases of liver cancer were identified in women aged 19-54 years; five controls selected from among patients hospitalized for acute conditions unrelated to oral contraceptive use were matched to each case on age (five-year categories), date of interview (three-year categories), and geographic location of the hospital. Among nine cases classified as having hepatocellular carcinoma, eight (89%) had used oral contraceptives; only 16 (36%) of 45 matched controls had used oral contraceptives. Among three other cases (two with cholangiocarcinomas and one with liver cancer of undetermined type), all had used oral contraceptives, compared with four of 15 matched controls. The results confirm the strong positive association between oral contraceptive use and hepatocellular carcinoma observed in earlier studies. Such an association is consistent with evidence that oral contraceptive use is associated with benign hepatic tumors in young women. However, the number of cases of liver cancer in the United States that are attributable to oral contraceptive use is probably small, because liver cancer is extremely rare in the United States.     

Fall-related occupational injuries on farms

We assessed risk factors for fall-related farm injuries in a population-based, case-control study. Cases had to reside in a defined geographic region served by a single medical center. Multiple sources reported cases, and a special farm census enabled random selection of controls. The annual risk of farm fall injury was 7.5 (95% CI: 5.7, 10.0) per 1,000 person-years. The crude incidence rate was higher in men, while the rate based on hours of farmwork was higher in women. In a multivariate analysis of risk factors, three factors were significantly associated with the risk. The risk of fall injury increased 2% (95% CI: 1%, 4%) per hour worked. Residents of farms with some farm workers not living on the farm had a fall injury rate 2.5 (95% CI: 1.0, 6.2) times greater than residents of other farms. Residents of farms with registered cows had one-third (95% CI: 0.14, 0.93) the risk of residents of other farms. To identify environmental hazards for fall injuries, researchers from several disciplines may need to collaborate in the design and conduct of studies that include injury site investigations. © 1996 Wiley-Liss, Inc.     

Efficient designs of gene-environment interaction studies: implications of Hardy-Weinberg equilibrium and gene-environment independence

It is important to investigate whether genetic susceptibility variants exercise the same effects in populations that are differentially exposed to environmental risk factors. Here, we assess the power of four two-stage case-control design strategies for assessing multiplicative gene-environment (G-E) interactions or for assessing genetic or environmental effects in the presence of G-E interactions. We considered a di-allelic single nucleotide polymorphism G and a binary environmental variable E under the constraints of G-E independence and Hardy-Weinberg equilibrium and used the Wald statistic for all tests. We concluded that (i) for testing G-E interactions or genetic effects in the presence of G-E interactions when data for E are fully available, it is preferable to ascertain data for G in a subsample of cases with similar numbers of exposed and unexposed and a random subsample of controls; and (ii) for testing G-E interactions or environmental effects in the presence of G-E interactions when data for G are fully available, it is preferable to ascertain data for E in a subsample of cases that has similar numbers for each genotype and a random subsample of controls. In addition, supplementing external control data to an existing case-control sample leads to improved power for assessing effects of G or E in the presence of G-E interactions. Copyright ? 2012 John Wiley & Sons, Ltd.     

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.     

Using dead controls to adjust for confounders in case-control studies

The use of dead controls in a case-control study, the primary purpose of which is to control for confounding, leads to consistent relative risk estimates within stratum of the confounder, even if the causes of death of controls are associated with that confounder. Dead controls may be more comparable to dead cases in terms of data quality, and therefore, under these circumstances, the use of dead controls may be advantageous.     

A case-control study found that low albumin and smoking were associated with aortic dissection

OBJECTIVE: Hypertension is a risk factor for aortic dissection. Besides hypertension, associated factors for incidence of aortic dissection are not clear. The aim of this study was to examine whether other factors exist by an epidemiologic case-control approach. STUDY DESIGN AND SETTING: 240 consecutive cases suffering from acute aortic dissection were hospitalized in our university from 1989 to 2000. Blood chemistries and other information obtained immediately after admissions for the first attack were available in 226 patients. A case-control approach was performed by frequency matching age, sex, and history of hypertension in the remaining 226 cases with those of 226 healthy controls drawn from a similar geographic catchment area and all observed in 1989. RESULTS: The mean albumin (3.6 +/- 0.5 vs. 4.3 +/- 0.2 g/dL: P<.001) level of cases was significantly lower than that of controls (odds ratio: 0.004 per g/dL). The frequency with which cases of current smokers was significantly higher than that of controls (odds ratio: 3.475). The above parameters of cases compared with controls were statistically significant after adjustments for age, sex, and history of hypertension. CONCLUSION: Low albumin levels and smoking in addition to hypertension are significantly associated with aortic dissection.     

Designing and analysing case-control studies to exploit independence of genotype and exposure

Genetic susceptibility and environmental exposures play a synergistic role in the aetiology of many diseases. We consider a case-control study of a rare disease in relation to a categorical exposure and a genetic factor under the assumption that the genotype and the exposure occur independently in the population under study. Using a logistic model for risk, we describe maximum likelihood methods based on log-linear models that explicitly impose the independence assumption, something the usual logistic regression analyses cannot do. The estimator of the genotype–exposure interaction effect depends only on data from cases. Estimators for genotype and for exposure effects depend also on data from controls, but only through their respective marginal totals. All three estimators have smaller variance than they would were independence not enforced. These results have important implications for design: (i) Case-only studies can efficiently estimate gene-by-environment interactions. (ii) Studies where controls are genotyped anonymously can estimate genotype, exposure, and interaction effects as efficiently as designs where genotype and exposure data are linked. This feature addresses a growing concern of human subjects review boards. (iii) Exposure and interaction effects, but not genotype effects, can be estimated from studies where genetic information is only collected from cases (although one can recover the genotype effect if external gene prevalence data exist). Such designs have the compensatory benefit that the response rate (hence, validity) is higher when controls are not subjected to intrusive tissue sampling. However, the independence assumption can be checked only with linked genotype and exposure data for some controls. We illustrate the methods by applying them to recent study of cleft palate in relation to maternal cigarette smoking and to a variant of the transforming growth factor alpha gene in the child. © 1997 by John Wiley & Sons, Ltd.     

Do cases and controls matched on the first eight digits of their telephone number share geographic proximity and socioeconomic characteristics?

PURPOSE: The aim of this study is to assess the geographic proximity of cases and controls who share the same "stem" of their telephone number (area code, prefix, and next two digits) and compare their socioeconomic characteristics. METHODS: We compared residential proximity and characteristics of case-control sets selected in this way with sets consisting of the same cases matched to different controls on all variables except telephone number stem. RESULTS: Mean distance between telephone stem-matched pairs was much less than that between pairs not matched by telephone number stem (3.53 and 20.14 km, respectively). There was slight agreement between cases and controls for some measures of socioeconomic status in both sets. We also compared distances between residences of control respondents who shared the same telephone number stem with those between residences of respondents who shared only the same telephone area code and prefix. Mean distances for the two groups were similar. CONCLUSIONS: Geographic proximity between cases and controls can be achieved by matching on either telephone number stem or only telephone area code and prefix. Adjustment for socioeconomic factors may be needed even if cases and controls live close to each other.