Marginal structural model to evaluate the association between cumulative osteoporosis medication and infection using claims data.

Due to the suboptimal persistence to osteoporosis (OP) treatment, factors triggering treatment discontinuation/switching may be causing time-varying confounding. BP treatment was associated with the risk of overall infection in opposite directions in the unweighted Cox model versus the weighted MSM. The discrepancy of effect estimates for overall infection in the MSM suggested there may be time-varying confounding. INTRODUCTION: Due to the suboptimal persistence to osteoporosis (OP) treatment, factors triggering treatment discontinuation/switching may be affected by prior treatment and confound the subsequent treatment effect, causing time-varying confounding. METHODS: In a US insurance database, the association between joint treatment of bisphosphonates (BP) and other OP medication and the incidence of infections among postmenopausal women was assessed using a marginal structural model (MSM). Stabilized weights were estimated by modeling treatment and censoring processes conditioning on past treatment, and baseline and time-varying covariates. RESULTS: BP treatment was associated with the risk of overall infection in opposite directions in the unweighted Cox model {incidence rate ratio [IRR] [95% confidence interval (CI)] = 1.15 [1.14-1.17]} versus the weighted MSM [IRR (95% CI) = 0.79 (0.77-0.81)], but was consistently associated with a lower risk of serious infection in both the unweighted Cox model [IRR (95% CI] = 0.79 (0.78-0.81)) and the weighted MSM [IRR (95% CI) = 0.71 (0.68-0.75)]. Similar results were found when current and past treatments were simultaneously assessed. CONCLUSIONS: The discrepancy of effect estimates for overall but not serious infection comparing unweighted models and MSM suggested analyses of composite outcomes with a wide range of disease severity may be more susceptible to time-varying confounding.

Study of adverse outcomes in women using testosterone therapy.

OBJECTIVES: There are concerns that exogenous testosterone therapy may be associated with adverse cardiovascular effects, increases in risk of breast or uterus cancer and alterations in insulin sensitivity. Objective of this study was to explore the safety of testosterone therapy in actual clinical practice. METHODS: Data from the General Practice Research Database and the Health Improvement Network was used, including computerised medical records of UK general practitioners. The study population included women aged 18+ years prescribed testosterone, administered through implants (72.2%), tablets (18.4%) or injections (7.9%). Each testosterone user was matched by age and practice to three control patients. Cox proportional hazards models were used to compare the rates of several outcomes. RESULTS: The study population included 8412 women, 2103 testosterone users and 6309 controls. There were no statistically significant differences between the cohorts in the rates of cerebrovascular disease, ischemic heart disease, breast cancer, deep venous thrombosis/pulmonary embolism, diabetes mellitus or acute hepatitis. The rate of breast cancer was comparable between testosterone users and control patients. The rate of androgenic events was increased in the testosterone cohort (relative rate of 1.55 [95% CI 1.21-1.97]). Differences in outcomes between the cohorts were generally comparable across subgroups based on age and use of hormone therapy. CONCLUSIONS: This study found no major increase in the risk of cardiovascular diseases or breast cancer in women using testosterone (implants, tablets, or injections), while the risk of androgenic events was increased. It would be useful to conduct similar studies at lower doses with transdermal testosterone.

Incidence and natural history of Paget's disease of bone in England and Wales.

This study used a large, primary care, record-linkage resource (the General Practice Research Database [GPRD]) to evaluate the incidence, clinical presentation, and natural history of Paget's disease of bone in England and Wales. Between 1988 and 1999, we identified 2465 patients with the recorded diagnosis of Paget's disease of bone, within the five million subjects > or = 18 years old who were registered in the GPRD. The validity of diagnostic recording was assessed by questionnaire to individual general practitioners (GPs) in 150 patients; the diagnosis was confirmed in 93.8% of responders. The mean age of patients with Paget's disease was 75 years and 51% were men. The prevalence of the disorder was 0.3% among men and women aged > or = 55 years; incidence rates for clinically diagnosed Paget's disease rose steeply with age (men, 5 per 10,000 person-years; women, 3 per 10,000 person-years at the age of 75 years). Over the 11-year period of the study, the age- and sex-adjusted incidence rate of clinically diagnosed Paget's disease declined from 1.1 per 10,000 person-years to 0.7 per 10,000 person-years. Each patient with Paget's disease was matched to three controls matched by age, gender, and general practice. Cases had a greater risk of back pain (relative risk [RR], 2.1; 95% CI, 1.9-2.3), osteoarthritis (OA; RR, 1.7; 95% CI, 1.5-1.9), hip arthroplasty (RR, 3.1; 95% CI, 2.4-4.1), knee arthroplasty (RR, 1.6; 95% CI, 1.0-2.6), fracture (RR, 1.2; 95% CI, 1.0-1.5), and hearing loss (RR, 1.6; 95% CI, 1.3-1.9). Seven patients with Paget's disease developed a malignant bone neoplasm (0.3%). Using life table methodology, the estimated number of people who died within 5 years of follow-up was 32.7% among the patients with Paget's disease and 28.0% among the control patients.

Family risk score of coronary heart disease (CHD) as a predictor of CHD: the Atherosclerosis Risk in Communities (ARIC) study and the NHLBI family heart study.

Family history of coronary heart disease (CHD) has been found to be a risk factor for CHD in numerous studies. Few studies have addressed whether a quantitative measure of family history of CHD (family risk score, FRS) predicts CHD in African Americans. This study assessed the association between FRS and incident CHD of participants, and the variation of the association by gender and race. Participants in the study were a biracial population-based cohort with 3,958 African Americans and 10,580 Whites aged 45-64 years old in the ARIC baseline survey (1987-1989). They were randomly selected from four U. S. communities. During follow-up (1987-1993), 352 participants experienced the onset of CHD. Incidence density of CHD (per 1,000 person-years) was 7.8 and 3.6 among African-American men (AAM) and women (AAW), and 7.2 and 2.2 among White men (WM) and women (WW). The hazard rate ratio (HRR) of CHD associated with one standard deviation increase of FRS was 1.52 in AAW, 1.46 in AAM, 1.41 in WW, and 1.68 in WM. The HRRs decreased 4.6% in AAW, 1.4% in WW, 5.7% in AAM, and 3.0% in WM, but increased 2.1% in AAM after adjustment for selected covariates. FRS predicts incident CHD in African Americans and Whites, men and women. The relation of FRS to incident CHD can be only partially explained by the selected risk factors in the biological causal pathways: IMT, T-G, LDL, HDL, Lp(a), fibrinogen and hypertension. No significant difference by race has been found in this study.

Trends in cigarette smoking: the Minnesota Heart Survey, 1980 through 1992.

OBJECTIVES: The purpose of this study was to describe trends in the prevalence of cigarette smoking between 1980 through 1982 and 1990 through 1992 in Minneapolis and St. Paul, Minn. METHODS: Three population-based surveys were conducted among adults 25 to 74 years of age in 1980 through 1982, 1985 through 1987, and 1990 through 1992. RESULTS: Overall age-adjusted prevalences of cigarette smoking declined significantly between 1980-1982 and 1985-1987 and between 1985-1987 and 1990-1992. Serum thiocyanate, a biochemical marker for tobacco use, also declined significantly over the 3 periods. CONCLUSIONS: Favorable trends in smoking prevalence and cigarette consumption among smokers were observed, but disturbing trends in some smoking behaviors were also noted.

Risk factors for kidney damage in the adult population of Wadena, Minnesota. A prospective study.

An increased albumin excretion rate (AER) is associated with impaired glucose tolerance and diabetes mellitus in some populations, but data on Americans of Northern European origin are lacking. In 1986-1987, AER and creatinine clearance were measured in 455 adults in a survey of the population of Wadena, Minnesota. Thirty-five subjects (8%) had an AER > or = 15 micrograms/minute, and eight of these had overt proteinuria (AER > or = 175 micrograms/minute). AER and creatinine clearance were uncorrelated except when AER was increased. Unadjusted mean AER in a stratified random sample of adults (n = 374) was 3.6 micrograms/minute. Adjusted values for 277 subjects with normal glucose tolerance and for 80 subjects with impaired glucose tolerance were very similar (3.8 and 3.7 micrograms/minute, respectively), whereas mean AER was 5.4 micrograms/minute for persons with non-insulin-dependent diabetes mellitus (NIDDM) who were not taking insulin and 9.4 micrograms/minute for persons with NIDDM who were taking insulin (p < 0.0001). After adjustment for age, mean creatinine clearance was unrelated to glucose tolerance. Systolic blood pressure was a major determinant of increased AER (p < 0.0001) and lowered creatinine clearance (p = 0.0011), independently of diabetes. AER was stable over 5 years among the 321 cases who were not taking insulin and were not severely hypertensive. The decrease in creatinine clearance was greater in ex-smokers and current smokers than in nonsmokers. The authors conclude that hypertension and NIDDM were independently associated with the risk of kidney damage in this population, as indicated by a higher AER. High-normal blood pressure, but not impaired glucose tolerance, was associated with microalbuminuria. These relatively mild changes may reflect an ethnically based resistance to the damaging effects of hyperglycemia on the kidney. Smoking may accelerate the aging-related decline in glomerular filtration rate.

Alcohol-containing mouthwashes and oropharyngeal cancer: a spurious association due to underascertainment of confounders?

Recently it has been suggested that the use of alcohol-containing mouthwashes may increase the risk of oropharyngeal cancer. Heavy alcohol intake and tobacco use are established causes of oropharyngeal cancer. Their use is associated with mouthwash use. In addition, alcohol and tobacco use both tend to be underreported. Here the authors show that, under the hypothesis that mouthwash does not increase the risk of oropharyngeal cancer, confounding due to underascertained exposure to alcohol and tobacco would result in a spuriously elevated odds ratio for mouthwash use. As a general principle, a null association becomes apparently positive if a confounding variable is incompletely ascertained: a spurious association may be produced even in the absence of a difference in the extent of the underascertainment of the confounder among the comparison groups.

NHLBI Family Heart Study: objectives and design.

The NHLBI Family Heart Study is a multicenter, population-based study of genetic and nongenetic determinants of coronary heart disease (CHD), atherosclerosis, and cardiovascular risk factors. In phase I, 2,000 randomly selected participants and 2,000 with family histories of CHD were identified among 14,592 middle-aged participants in epidemiologic studies. Medical histories from these individuals, their parents, and their siblings were used to calculate family risk scores that compared the number of reported and validated CHD events with the number expected based on the size, sex, and age of family members. A total of 657 families with the highest risk scores and early-onset CHD and 588 randomly sampled families had clinic examinations that included electrocardiograms, carotid artery ultrasound scans, spirometry, measurements of body size, blood pressure, lipids, lipoproteins, hemostatic factors, insulin, glucose, and routine chemistries. Additional biochemical and genetic studies are being performed on selected participants. Serum, plasma, lymphocytes, red cells, and DNA are stored for future studies, including genotyping of candidate genes and anonymous markers. Contributions of genes, shared and individual environments, and behaviors to variations in risk factors, preclinical atherosclerosis, and CHD will be estimated. Linkage studies, including the quantitative trait loci approach, are planned.

Sex differences in early mortality after acute myocardial infarction (the Minnesota Heart Survey).

Although numerous studies indicate that women have a higher early mortality from acute myocardial infarction (AMI) than men, reasons for the difference are largely unexplained. We studied the role of sex in the prognosis of 1,600 patients with AMI aged 30 to 74 years in the population-based Minnesota Heart Survey. A 50% random sample was taken of all AMI patients hospitalized in 1980 and 1985 in the Twin Cities of Minnesota (Minneapolis-St. Paul) (1,168 men, 432 women). A multiple logistic regression model was used for predicting early death (within 28 days) and included baseline characteristics: sex, age, chest pain on admission, history of previous AMI, angina pectoris, coronary artery bypass surgery or hypertension, presence of heart failure, cardiac arrhythmias requiring direct-current shock, diabetes mellitus, valvular disease, cardiomyopathy, and levels of serum enzymes and blood urea nitrogen. Age-adjusted early mortality rate was significantly higher in women than men, but only in those aged < 65 years (12.5% of women vs 6.5% of men, p < 0.01) versus those aged > or = 65 years (19.5% vs 21.6%, p > 0.05). Multivariate analysis also showed that among those < 65 years, female sex was a strong and independent predictor of early death (odds ratio 2.0, 95% confidence interval 1.2 to 3.5, p < 0.01). Rates of coronary angiography, coronary artery bypass surgery, percutaneous transluminal coronary angioplasty, and thrombolysis performed during hospital stay were higher in men, but after adjustment for age, congestive heart failure, and diabetes mellitus, a statistically significant difference persisted only in the frequency of coronary angiography (26% in men vs 17% in women, p < 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)

Validation of self-reported history of acute myocardial infarction: experience of the Minnesota Heart Survey Registry.

Accurate separation of new cases of acute myocardial infarction from prevalent cases is critical for assessing trends in morbidity in population-based studies. This report presents data on the validity of self-reported history of previous acute myocardial infarction among 3,703 patients admitted to a coronary care unit with suspicion of acute myocardial infarction. We substantiated the history of a prior event for 60% of those who reported one (629 of 1,053) and found 40% to be false-positive histories. Much of the false-positive reporting was related to previous cardiac hospitalizations, predominantly (40%) for unstable angina.

Improved survival of stroke patients during the 1980s. The Minnesota Stroke Survey.

BACKGROUND AND PURPOSE: The underlying reasons for the decline in stroke mortality in the United States are not well understood and have been the subject of ongoing debate. This study was undertaken to determine whether survival of hospitalized stroke patients has changed during the 1980s, thereby contributing to the decline in stroke mortality during that period. METHODS: For the years 1980, 1985, and 1990, we obtained listings of discharge diagnoses from hospitals in the Minneapolis-St Paul metropolitan area and identified all hospitalizations with a discharge diagnosis code of acute cerebrovascular disease according to the International Classification of Diseases, 9th Revision. A 50% random sample of men and women aged 30 to 74 years was selected in each survey for detailed medical record abstraction. Standardized sets of criteria for stroke were then used to validate acute stroke events throughout the 1980s. Each of the three period cohorts of hospitalized stroke patients (1980, 1985, and 1990) was followed for at least 2 years for all-cause mortality end point. RESULTS: A total of 1853 patients met minimal criteria for acute stroke: 564 patients in 1980, 598 patients in 1985, and 691 patients in 1990. Controlling for age, the odds of death within 2 years after stroke were approximately 40% lower in 1990 than in 1980. The relative odds of 2-year death in 1990 (versus 1980) were 0.65 (95% confidence interval, 0.47 to 0.89) and 0.60 (95% confidence interval, 0.42 to 0.85) for men and women, respectively. The improved survival was evident in the short term (28 days) as well as for stroke patients who survived that period. Analysis according to stroke subtype revealed that improved survival of ischemic stroke and specifically of stroke with no apparent cardioembolic source largely accounted for the overall trend. The prognosis of stroke patients who were admitted in a comatose state has not changed during that decade. CONCLUSIONS: Despite the absence of any clear major advances in acute stroke therapy, survival of stroke patients substantially improved during the 1980s. The underlying reasons for this unexpected yet remarkable trend remain uncertain but may include improved supportive and rehabilitative care of stroke victims as well as a change in the natural history of the disease.

A reassessment of fasting plasma glucose concentrations in population screening for diabetes mellitus in a community of northern European ancestry: the Wadena City Health Study.

In current clinical and research practice, the determination of diabetic status depends largely on plasma glucose levels 2 h after the ingestion of a standard 75-g glucose load, the oral glucose tolerance test (OGTT). The OGTT, however, remains inconvenient, not highly reproducible, and costly, especially for large-scale studies and population screening tests. Fasting plasma glucose (FPG) determinations are convenient, reliable, and valid measures of glucose intolerance, but the currently prescribed cut-off point of 140 mg/dl (7.8 mM) lacks sensitivity. We evaluated the reliability and validity of fasting plasma glucose (FPG) values compared with other measures of hyperglycemia for a diagnosis of diabetes in a population-based study of carbohydrate metabolism in Wadena, Minnesota, a community of predominantly northern European ancestry. As a part of this effort, a random sample of Wadena adults, stratified by age and gender, plus all known, previously diagnosed diabetics participated in 2 days of baseline testing and were followed prospectively and retested 5 years later. Cross-sectional analyses of baseline data are presented in this article. Diabetic status was ascertained by administering a standard OGTT according to National Diabetes Data Group (NDDG) specifications. Sensitivity and specificity levels obtained when using a FPG cut-off point of 6.4 mM were 95.2% and 97.4%, respectively. In study subjects with no known diagnosis of diabetes, the FPG cut-off point of 6.4 mM performed reasonably well with a sensitivity and specificity of 67.7% and 97.4%, respectively. (ABSTRACT TRUNCATED AT 250 WORDS)

Predictors of cancer prevention attitudes and participation in cancer screening examinations.

BACKGROUND: Few current data are available regarding factors associated with participation in cancer screening examinations in the general population. METHODS: To identify factors associated with participation in cancer screening examinations, random population samples of 25- to 74-year-old men and women in six various-sized communities in three upper-Midwestern states (n = 4,915) were surveyed in 1987-1989. Multivariate-adjusted means were calculated and compared using analysis of covariance. RESULTS: Statistically significant (P < 0.05) strong predictors (other than age and sex) of ever having had a specific cancer screening test were as follows (the numbers in parentheses following each listed association are the absolute maximum differences in mean proportions among the levels of the predictors): (1) rectal examination: higher education (14%); (2) fecal occult blood testing: higher education (6%) and never smoker (5%); (3) sigmoidoscopy: higher income (7%) and higher education (6%); and (5) mammography: higher income (25%), higher education (8%), and a positive family history of breast cancer (7%). There were no strong predictors (out of nine) of ever having had a Papanicolaou smear or a breast self-examination. CONCLUSIONS: The largest differences among the population for participation in cancer screening examinations involves income and the two most expensive cancer screening tests: higher income is a strong predictor of having a mammogram and, to a lesser extent, of having a sigmoidoscopy. The most consistent predictor of participation in cancer screening examinations across all cancer screening tests is education: higher education is a predictor of having each kind of cancer screening test.

Survey of health and use characterization of pesticide appliers in Minnesota.

We surveyed 1,000 randomly selected state-licensed pesticide appliers to improve our understanding of pesticide use and its potential health effects. Participants were stratified by pesticide class (herbicides, insecticides, fungicides, fumigants) to determine potential differences in health characteristics among different pesticide groups. A subset of 60 applicators, divided by pesticide class used, were studied for exposure-related cholinesterase (ChE) depression. ChE depression in excess of 20% was most frequent in fumigant applicators who did enclosed-space application, in addition to other pesticide application procedures (p < .05). Survey data demonstrated that the prevalence of all common chronic diseases considered together was significantly increased (p = .015) in fumigant appliers, compared with all other pesticide use groups. The frequency of chronic lung disease was also significantly increased in the fumigant applier group (p = .027). Curiously, two cases of a rare hematopoietic neoplasm--hairy cell leukemia--were identified in our study group (annual incidence 0.67/100,000 in Minnesota). Whether there is an association between this unique tumor and agricultural work is uncertain, and further study is needed in this regard.

Trends in coded causes of death following definite myocardial infarction and the role of competing risks: the Minnesota Heart Survey (MHS).

We investigated possible differences over time in underlying causes of death among validated definite myocardial infarction cases who were discharged following an index hospitalization in 1970, 1980, and 1985 in the Twin Cities, MN. No changes were observed in underlying causes of death assigned to patients who died prior to discharge in the 3 years. Among in-hospital survivors of definite MI, however, age-adjusted rates of death from non-cardiovascular causes more than doubled between 1970 and 1985 (P < 0.01). More specifically, mortality rates for diabetes mellitus increased significantly from 1970 to 1985 (P < 0.05), while those for neoplasms and diseases of the respiratory system increased non-significantly. Whether these data are the result of artifactual changes in cause of death assignment or real changes in disease severity and comorbidity, these trends in long-term death following acute MI may have had a modest impact on reported community-wide coronary heart disease mortality rates.

Apparently coronary heart disease-free patients in the coronary care unit: characteristics, medical care, and 1-year outcome.

BACKGROUND: Coronary care units (CCUs) have contributed significantly to the improved survival rates among patients with acute myocardial infarction. Many patients admitted to CCUs are certified to be free of coronary heart disease (CHD) at discharge. There is little literature on the hospital course and prognosis of such patients. METHODS: We identified and followed 594 patients admitted to six CCUs in the Minneapolis-St Paul metropolitan area in 1990 because of suspected acute myocardial infarction who were eventually discharged without evidence of acute or chronic CHD. Their baseline characteristics, medical care, and 1-year outcome were compared with those of 672 patients with confirmed acute myocardial infarction and 612 patients with a history of CHD but without evidence of an acute coronary event. RESULTS: Similar numbers of men and women were certified to be CHD-free on discharge from hospital. These patients were significantly younger than either patients with acute myocardial infarction or patients with a history of CHD (mean age 57, 65, and 67 years, respectively). CHD-free patients commonly reported current smoking, hypertension, and hypercholesterolemia (26, 50, and 18%, respectively). These patients were less likely than those with acute myocardial infarction or a history of CHD to undergo diagnostic or therapeutic procedures, or to receive pharmacological treatment. Their 1-year mortality rate was 5%, significantly lower (P < 0.05) than the mortality among patients with either acute myocardial infarction (18%) or a history of CHD (13%) but 2.6 times greater than expected in the general population. Older age, previous or current smoking, chest pain leading to admission, and congestive heart failure were independent predictors of 1-year mortality. CONCLUSIONS: Patients certified to be CHD-free after admission to a CCU with suspected acute myocardial infarction have a lower 1-year mortality rate than patients experiencing acute myocardial infarction or chronic CHD. Their mortality rate, however, is substantially higher than expected, probably because of a high prevalence of cigarette smoking and hypertension.

Trends in diabetes prevalence among stroke patients and the effect of diabetes on stroke survival: the Minnesota Heart Survey.

This study documented trends in the prevalence of diabetes among men and women hospitalized for acute stroke and determined the effect of diabetes on short- and long-term survival following stroke. These issues were investigated in the Minnesota Heart Survey, a population-based surveillance system that has monitored trends in stroke morbidity and mortality in the Minneapolis-St Paul metropolitan area since 1970. Clinical data were obtained from the hospital records of 50% samples of residents ages 30 to 74 years who were discharged with a diagnosis of acute stroke in 1970, 1980, and 1985. Between 1970 and 1985, the prevalence of diabetes as listed on the discharge diagnoses among stroke patients increased significantly in men (22.4% vs 10.5%; p = 0.006) and non-significantly in women (24.7% vs 15.9%; p = 0.3). During this time period, both in-hospital and 28-day case fatality rates declined in non-diabetic stroke patients but remained unchanged in stroke patients with diabetes. After controlling for the effects of age, sex, survey year, and level of consciousness, diabetes status had little effect on short-term (28-day) mortality of stroke patients, but the odds of 5-year mortality among those surviving to 1 year was 2.0 (95% Cl (1.3, 3.2)) times higher in diabetic compared to non-diabetic individuals. These findings suggest that the prevalence of diabetes has been increasing among stroke patients, and that the diabetic condition is a significant predictor of poorer long-term but not short-term survival following stroke.

Relationship between flexible sigmoidoscopy training during residency and subsequent sigmoidoscopy performance in practice.

BACKGROUND: The primary objective of the research was to determine whether sigmoidoscopy training during family practice residency is associated with subsequent performance of sigmoidoscopy in practice. METHODS: We surveyed 292 family physicians who graduated from the University of Minnesota Department of Family Practice and Community Health residency program between 1983 and 1989. The survey instrument collected information on the number of physicians who were currently performing flexible sigmoidoscopy in their practices and what factors were associated with performance of this procedure. RESULTS: Physicians with flexible sigmoidoscopy training during residency were performing flexible sigmoidoscopies at a significantly higher rate than those without training during residency (P = .001). A significantly higher proportion of males were performing flexible sigmoidoscopy in their practices than females (P = .0002). The mean number of flexible sigmoidoscopies recommended by residency-trained physicians to be performed during residency for adequate training was 16. CONCLUSIONS: Training in flexible sigmoidoscopy during a family practice residency is associated with a higher rate of flexible sigmoidoscopy performance later in practice. Female physicians perform flexible sigmoidoscopy at a significantly lower rate than their male colleagues; this could be due to a less-adequate training during residency. We recommend that residents perform a minimum of 16 flexible sigmoidoscopies during residency training.

Does body fat distribution promote familial aggregation of adult onset diabetes mellitus and postmenopausal breast cancer?

Noninsulin-dependent diabetes mellitus and postmenopausal breast cancer share a number of risk factors, including obesity, increased waist-to-hip ratio, and a positive family history. If risk for these diseases is mediated through a familial tendency for abdominal obesity, then one might expect to see familial clustering of both diseases. We analyzed data from a prospective cohort study of 41,837 Iowa women age 55-69 years. Diabetes was not associated with incidence of breast cancer [relative risk (RR) = 0.97]. The association between family history of breast cancer and breast cancer incidence, however, was slightly modified by individual history of diabetes: a positive family history of breast cancer in the absence of baseline diabetes was associated with a relative risk of 1.36 [95% confidence interval (CI) = 1.08-1.70], whereas the presence of both factors was associated with a RR of 1.87 (95% CI = 0.93-3.76). Adjustment for waist-to-hip ratio greatly diminished this difference. Conversely, a family history of breast cancer was associated with a RR of 5-year diabetes mortality of 1.94 (95% CI = 1.17-3.24) that persisted after stratification by tertile of waist-to-hip ratio. No clear association of family history of breast cancer and waist-to-hip ratio for self-reported diabetes incidence was evident. These data are indicative of a complex interrelation between waist-to-hip ratio, familial predisposition, diabetes, and breast cancer.