American Journal of Epidemiology - current issue

Cover

2009-06-24

Editorial Board

2009-06-24

A Field Synopsis and Meta-Analysis of Genetic Association Studies in Peripheral Arterial Disease: The CUMAGAS-PAD Database

Zintzaras, E., Zdoukopoulos, N. — 2009-06-24

In an electronic search of the literature, the authors systematically retrieved all published studies that investigated genetic susceptibility to peripheral arterial disease (PAD). They created a comprehensive database of all eligible studies, collecting detailed genetic and bioinformatics data on each polymorphism. Data from eligible studies were synthesized using meta-analysis techniques. Gene variants were classified into distinct pathophysiologic pathways, and their potential involvement in PAD pathogenesis was determined. Forty-one publications that examined 44 gene polymorphisms were included. For 37 polymorphisms, the variant form had a functional effect. Twenty-three polymorphisms in 22 potential PAD candidate genes (F2, FGB, MTHFR, ITGB3, ACE, AGT, IL6, CCL2, ICAM1, SELE, MMP9, PPARG, MMP1, ADD1, P2RY12, LIPC, PLA2G7, SCARB1, MMP3, MTTP, LPA, CHRNA3) showed a significant association in individual studies. Eighty-eight percent of the studies had statistical power of less than 50%, and in 15 studies the genotype distribution in the control group did not conform to Hardy-Weinberg equilibrium. Data on 12 polymorphisms (F5 1691 G/A, MTHFR 677C/T, F2 20210 G/A, ITGB3 1565 T/C, ACE I/D, AGT 704C/T, AGT -6G/A, AGT 733C/T, IL6 -174 G/C, MMP9 -1562C/T, ICAM1 1462A/G, CHRNA3 831C/T) were synthesized, and a positive association was found for 3 (IL6 -174 G/C, ICAM1 1462A/G, CHRNA3 831C/T).

Benefits and Risks of Postmenopausal Hormone Therapy When It Is Initiated Soon After Menopause

2009-06-24

The authors further analyzed results from the Women's Health Initiative randomized trials (1993–2004) of conjugated equine estrogens, with or without medroxyprogesterone acetate, focusing on health benefits versus risks among women who initiated hormone therapy soon after menopause. Data from the Women's Health Initiative observational study (1993–2004) were included in some analyses for additional precision. Results are presented here for incident coronary heart disease, stroke, venous thromboembolism, breast cancer, colorectal cancer, endometrial cancer, or hip fracture; death from other causes; a summary global index; total cancer; and total mortality. Hazard ratios for breast cancer and total cancer were comparatively higher (P < 0.05) among women who initiated hormone therapy soon after menopause, for both regimens. Among these women, use of conjugated equine estrogens appeared to produce elevations in venous thromboembolism and stroke and a reduction in hip fracture. Estrogen plus progestin results among women who initiated use soon after menopause were similar for venous thromboembolism, stroke, and hip fracture but also included evidence of longer-term elevations in breast cancer, total cancer, and the global index. These analyses provide little support for the hypothesis of favorable effects among women who initiate postmenopausal estrogen use soon after menopause, either for coronary heart disease or for health benefits versus risk indices considered.

Invited Commentary: Hormone Therapy Risks and Benefits--The Women's Health Initiative Findings and the Postmenopausal Estrogen Timing Hypothesis

Banks, E., Canfell, K. — 2009-06-24

Worldwide evidence on menopausal hormone therapy shows that it does not reduce coronary heart disease (CHD) risk and that it increases the risks of breast cancer, stroke, and venous thromboembolism. These risks are not offset by reductions in hip fracture risk. Consequently, the Food and Drug Administration and other drug regulatory authorities agree that hormone therapy should be used chiefly for short-term relief of menopausal symptoms. Continuing speculation relates to the "postmenopausal estrogen timing" hypothesis, which proposes that hormone therapy initiated soon after menopause will prevent CHD while therapy started later will have a null or adverse effect. The detailed analyses of Women's Health Initiative data reviewed here specifically address the timing hypothesis. For hormone therapy initiated soon after menopause versus therapy started later, the findings demonstrate 1) similar null or adverse effects on CHD risk; 2) similar adverse effects on the risks of stroke and venous thrombosis; and 3) possibly greater adverse effects on breast cancer risk. Therefore, Women's Health Initiative data do not support the hypothesis of favorable effects in women starting hormone therapy soon after menopause. Hence, the overall trial findings, including net harm for combined estrogen-progestin and the lack of a net benefit for estrogen-only therapy, also apply to women initiating hormone therapy soon after menopause.

Fast-Food Consumption, Diet Quality, and Neighborhood Exposure to Fast Food: The Multi-Ethnic Study of Atherosclerosis

Moore, L. V., Diez Roux, A. V., Nettleton, J. A., Jacobs, D. R., Franco, M. — 2009-06-24

The authors examined associations among fast-food consumption, diet, and neighborhood fast-food exposure by using 2000–2002 Multi-Ethnic Study of Atherosclerosis data. US participants (n = 5,633; aged 45–84 years) reported usual fast-food consumption (never, <1 time/week, or ≥1 times/week) and consumption near home (yes/no). Healthy diet was defined as scoring in the top quintile of the Alternate Healthy Eating Index or bottom quintile of a Western-type dietary pattern. Neighborhood fast-food exposure was measured by densities of fast-food outlets, participant report, and informant report. Separate logistic regression models were used to examine associations of fast-food consumption and diet; fast-food exposure and consumption near home; and fast-food exposure and diet adjusted for site, age, sex, race/ethnicity, education, and income. Those never eating fast food had a 2–3-times higher odds of having a healthy diet versus those eating fast food ≥1 times/week, depending on the dietary measure. For every standard deviation increase in fast-food exposure, the odds of consuming fast food near home increased 11%–61% and the odds of a healthy diet decreased 3%–17%, depending on the model. Results show that fast-food consumption and neighborhood fast-food exposure are associated with poorer diet. Interventions that reduce exposure to fast food and/or promote individual behavior change may be helpful.

Confounding by Dietary Patterns of the Inverse Association Between Alcohol Consumption and Type 2 Diabetes Risk

Imamura, F., Lichtenstein, A. H., Dallal, G. E., Meigs, J. B., Jacques, P. F. — 2009-06-24

The ability to interpret epidemiologic observations is limited because of potential residual confounding by correlated dietary components. Dietary pattern analyses by factor analysis or partial least squares may overcome the limitation. To examine confounding by dietary pattern as well as standard risk factors and selected nutrients, the authors modeled the longitudinal association between alcohol consumption and 7-year risk of type 2 diabetes mellitus in 2,879 healthy adults enrolled in the Framingham Offspring Study (1991–2001) by Cox proportional hazard models. After adjustment for standard risk factors, consumers of ≥9.0 drinks/week had a significantly lower risk of type 2 diabetes mellitus compared with abstainers (hazard ratio = 0.47, 95% confidence interval (CI): 0.27, 0.81). Adjustment for selected nutrients had little effect on the hazard ratio, whereas adjustment for dietary pattern variables by factor analysis significantly shifted the hazard ratio away from null (hazard ratio = 0.33, 95% CI: 0.17, 0.64) by 40.0% (95% CI: 16.8, 57.0; P = 0.002). Dietary pattern variables by partial least squares showed similar results. Therefore, the observed inverse association, consistent with past studies, was confounded by dietary patterns, and this confounding was not captured by individual nutrient adjustment. The data suggest that alcohol intake, not dietary patterns associated with alcohol intake, is responsible for the observed inverse association with type 2 diabetes mellitus risk.

Familial Recurrence of Midline Birth Defects--A Nationwide Danish Cohort Study

Oyen, N., Boyd, H. A., Poulsen, G., Wohlfahrt, J., Melbye, M. — 2009-06-24

If birth defects resulting from fusion failure in the midline have a common etiology, as previously hypothesized, persons with a family history of 1 type of midline defect should have an increased risk of dissimilar midline defects. The authors examined this hypothesis by linking information from the National Patient Register, the Causes of Death Register, and the Danish Family Relations Database for all Danish residents registered in the Civil Registration System during 1977–2005. Linkage yielded a cohort of 1.7 million persons with 1 or more relatives, including 9,063 persons with 1 or more midline defects. The authors investigated familial clustering of midline defects by estimating relative risks of similar and dissimilar midline defects according to family history of midline defects. Given a history of similar defects in first-degree relatives, relative risks for neural tube defects, conotruncal defects, oral facial clefts, anal/rectal defects, and diaphragmatic defects were 8.2 (95% confidence interval (CI): 3.1, 21.7), 7.7 (95% CI: 4.3, 13.8), 13.2 (95% CI: 10.8, 16.2), 10.3 (95% CI: 2.6, 41.1), and 11.2 (95% CI: 1.6, 79.7), respectively. However, given a dissimilar defect in a family member, the relative risk for any midline defect was null. Thus, similar defects but not dissimilar defects clustered in families, providing no evidence of a shared etiology for dissimilar midline defects.

Height and Site-specific Cancer Risk: A Cohort Study of a Korean Adult Population

Sung, J., Song, Y.-M., Lawlor, D. A., Smith, G. D., Ebrahim, S. — 2009-06-24

To evaluate the association between height and risk of cancer in an East Asian, middle-income population, the authors followed up a cohort of 788,789 Koreans (449,214 men and 339,575 women) aged 40–64 years for cancer incidence between 1994 and 2003. Cox proportional hazards regression analysis was used to evaluate the association. Each 5-cm increment in height was associated with 5% and 7% higher risk of all-sites cancer in men and women, respectively, after adjustment for age, body mass index, and behavioral and socioeconomic factors. When the associations were evaluated for site-specific cancers, a positive association was observed for cancer of the colon and thyroid in both men and women. Among gender-specific cancers, prostate cancer was positively associated with height in men. In women, there was a positive association between height and cancers of the breast and ovary, which did not change even after additional adjustment for reproductive factors. Although more clarification is needed for some site-specific cancers, the same positive association of height with cancer in a middle-income Korean population as found in high-income Western populations supports the influence of early life environment on cancer development in adulthood.

Impact of X Chromosome Genes in Explaining the Excess Risk of Cancer in Males

Biggar, R. J., Bergen, A. W., Poulsen, G. N. — 2009-06-24

The authors examined cancer incidence sex ratios in Denmark for 1943–2003 by age group. At nongenital/nonbreast sites, incidences were consistently higher in males. While environmental factors dominate cancer risk, the authors hypothesized that the higher risk in males might be explained by unspecified X chromosome genes’ protecting female cells from genotoxic damage. If so, cancer susceptibility would be passed from parent to offspring differently by sex. The authors compared relative risks in offspring of parents with and without cancer histories. For all comparisons, relative risks were similar in offspring of fathers with cancer (relative risk (RR) = 1.14, 95% confidence interval: 1.08, 1.20). Risks in offspring were higher for parents diagnosed before age 50 years and for cancers at the same site rather than different sites. Genital cancer risks were increased in same-sex offspring of parents with genital cancers. Breast cancer risks were high in both daughters (RR = 2.37) and sons (RR = 4.63) of mothers with breast cancer and in daughters (RR = 5.96) of fathers with breast cancer. Thus, X chromosome genetic factors were not responsible for the excess risk of cancer in males. Susceptibility to genital cancer was increased in same-sex offspring, and breast cancer risks were increased in both sons and daughters when either parent had had breast cancer.

Birth Weight, Postnatal Growth, and Age at Menarche

Terry, M. B., Ferris, J. S., Tehranifar, P., Wei, Y., Flom, J. D. — 2009-06-24

Larger body size in childhood is correlated with earlier age at menarche; whether birth and infant body size changes are also associated with age at menarche is less clear. The authors contacted female participants enrolled in the New York site of the US National Collaborative Perinatal Project born between 1959 and 1963 (n = 262). This racially and ethnically diverse cohort (38% white, 40% African American, and 22% Puerto Rican) was used to investigate whether maternal (body size, pregnancy weight gain, age at menarche, smoking) and birth (birth weight, birth length, placental weight) variables and early infant body size changes were associated with age at menarche even after considering later childhood body size. Higher percentile change in weight from ages 4 months to 1 year was associated with earlier age at menarche even after adjustment for later childhood growth (β = –0.15, 95% confidence interval: –0.27, –0.02 years per 10-percentile change in weight from ages 4 months to 1 year). The association was in the same direction for all 3 racial/ethnic groups but was largest for the white group. These New York Women's Birth Cohort Adult Follow-up data (2001–2006) suggest that infant weight gain, in addition to childhood weight gain, may be associated with earlier age at menarche.

Tracking Population Health Based on Self-reported Impairments: Trends in the Prevalence of Hearing Loss in US Adults, 1976-2006

Ikeda, N., Murray, C. J. L., Salomon, J. A. — 2009-06-24

Trends in the prevalence of hearing loss among US adults remain ambiguous because of variation across surveys in question wording and limited use of audiometric examinations. Pooling samples of participants aged 20–69 years in 4 nationally representative cross-sectional survey series conducted from 1976 to 2006 (N = 990,609), the authors performed logistic regression to quantify self-reporting biases compared with audiometric measurements. Statistically significant underreporting or overreporting of hearing loss was observed, with various patterns of bias across age groups and surveys. Substantial upward reporting biases appeared among young adults in the National Health and Nutrition Examination Survey since 1999 and in the National Health Interview Survey since 1997. Trends in age-standardized prevalence of bilateral hearing loss were estimated with corrections for self-reporting biases. Prevalence in men shifted from 9.6% (95% confidence interval (CI): 7.8, 11.8) in 1978 to 12.2% (95% CI: 10.1, 14.7) in 1993 and declined to 8.1% (95% CI: 7.0, 9.5) in 2000. In women, prevalence was relatively constant at approximately 6%–7% until the early 1990s and decreased from 7.0% (95% CI: 5.5, 9.1) in 1993 to 4.2% (95% CI: 3.4, 5.3) in 2000. Prevalence was stable in both sexes in the early 2000s. This approach to adjust for biases in self-reported impairments by using measured performance may be useful in various health domains.

Comparing Methods for Accounting for Seasonal Variability in a Biomarker When Only a Single Sample Is Available: Insights From Simulations Based on Serum 25-Hydroxyvitamin D

2009-06-24

In biomarker-disease association studies, the long-term average level of a biomarker is often considered the optimal measure of exposure. Long-term average levels may not be accurately measured from a single sample, however, because of systematic temporal variation. For example, serum 25-hydroxyvitamin D (25(OH)D) concentrations may fluctuate because of seasonal variation in sun exposure. Association studies of 25(OH)D and cancer risk have used different strategies to minimize bias from such seasonal variation, including adjusting for date of sample collection (DOSC), often after matching on DOSC, and/or using season-specific cutpoints to assign subjects to exposure categories. To evaluate and understand the impact of such strategies on potential bias, the authors simulated a population in which 25(OH)D levels varied between individuals and by season, and disease risk was determined by long-term average 25(OH)D. Ignoring temporal variation resulted in bias toward the null. When cutpoints that did not account for DOSC were used, adjustment for DOSC sometimes resulted in bias away from the null. Using season- or month-specific cutpoints reduced bias toward the null and did not cause bias away from the null. To avoid potential bias away from the null, using season- or month-specific cutpoints may be preferable to adjusting for DOSC.

Validity of Estimated Dietary Eicosapentaenoic Acid and Docosahexaenoic Acid Intakes Determined by Interviewer-Administered Food Frequency Questionnaire Among Older Adults With Mild-to-Moderate Cognitive Impairment or Dementia

2009-06-24

Epidemiologic research is increasingly being focused on elderly persons, many of whom exhibit mild-to-moderate cognitive impairment. This presents a challenge for collection and interpretation of self-reported dietary data. There are few reports on the impact of cognitive function and dementia on the validity of self-reported dietary intakes. Using plasma phospholipid fatty acid profiles as a biomarker of intake, the authors assessed the validity of an interviewer-administered food frequency questionnaire (FFQ) to estimate intakes of 2 marine-based omega-3 fatty acids, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), among 273 community-dwelling adults aged ≥60 years participating in the Nutrition, Aging, and Memory in Elders Study (Boston, Massachusetts, 2002–2008). Age- and energy-adjusted Pearson correlation coefficients for correlations between dietary intakes and plasma phospholipids were consistent across categories of high and low cognitive function (r = 0.48), based on Mini-Mental State Examination score, and were similar across clinically diagnosed categories of normal functioning (r = 0.49), mild cognitive impairment (r = 0.45), and dementia (r = 0.52). The FFQ ranked 78% of subjects to within 1 quartile of their plasma phospholipid EPA + DHA quartile. This frequency was consistently high across all cognitive categories. With interviewer administration, this FFQ seems to be a valid method of assessing dietary EPA + DHA intake in older adults with mild-to-moderate cognitive impairment.

Estimating Nutrient Intake From a Food Frequency Questionnaire: Incorporating the Elements of Race and Geographic Region

2009-06-24

Assignment of nutrient values to food frequency questionnaire (FFQ) items does not usually account for participant characteristics (besides age or sex) that may influence eating patterns. For the Southern Community Cohort Study, the authors developed and assessed results from a nutrient database system incorporating sex-, race-, and census-region-specific food lists, using 24-hour recall data from the National Health and Nutrition Examination Survey (NHANES III, NHANES 1999–2000, NHANES 2001–2002, and NHANES 2003–2004) and the Continuing Survey of Food Intakes by Individuals that permitted estimation of nutrients tailored to participants’ characteristics. For each of 15 nutrients, comparisons were made to a "standard" nutrient scoring system based on nationwide race-blind 24-hour recalls from these same sources. Using FFQ data from 67,926 Southern Community Cohort Study participants (47,038 African-American, 20,888 non-Hispanic white) aged 40–79 years who enrolled in the study during 2002–2008, the region- and race-informed system tended to produce increased estimated intake for most nutrients for black women, particularly for saturated fat (7.1%), monounsaturated fat (8.3%), and polyunsaturated fat (7.2%); smaller but significant changes (<5%) were also observed for nutrient intake for men and white women. These types of refinements in nutrient databases can be considered a means of enhancing the accuracy of dietary estimation using FFQs.

Methodological Issues in a Retrospective Cancer Incidence Study

2009-06-24

The authors traced incidence of central nervous system cancer in a large occupational cohort of jet engine manufacturing workers from 1976 to 2004 in the 24 US states that comprised 95% of the cohort deaths. The cohort of approximately 224,000 employees was matched with cancer registry data; all central nervous system cancer matches were requested with their diagnostic data. This paper highlights the obstacles encountered while conducting this retrospective cancer incidence study. The authors spent approximately 700 hours completing applications and obtaining the cohort matches. Approximately 70% of the cases were identified in the state in which the facility of interest is located. In addition to the large amount of time involved, identified issues include complicated approval processes, high costs, temporal differences among the registries, and registry agency difficulty in performing the matching. Several states do not allow individual-level data to be used for research purposes. Researchers can gain important cancer incidence information by matching retrospective cohorts to multiple state cancer registries. However, they should carefully weigh the time and costs required and plan accordingly. Despite some serious obstacles, many of which are potentially resolvable, cancer incidence studies of retrospective cohorts using multiple cancer registries are feasible.

The Nicaraguan Pediatric Dengue Cohort Study: Study Design, Methods, Use of Information Technology, and Extension to Other Infectious Diseases

2009-06-24

Dengue is a mosquito-borne viral disease that is a major public health problem worldwide. In 2004, the Pediatric Dengue Cohort Study was established in Managua, Nicaragua, to study the natural history and transmission of dengue in children. Here, the authors describe the study design, methods, and results from 2004 to 2008. Initially, 3,721 children 2–9 years of age were recruited through door-to-door visits. Each year, new children aged 2 years are enrolled in the study to maintain the age structure. Children are provided with medical care through the study, and data from each medical visit are recorded on systematic study forms. All participants presenting with suspected dengue or undifferentiated fever are tested for dengue by virologic, serologic, and molecular biologic assays. Yearly blood samples are collected to detect inapparent dengue virus infections. Numerous information and communications technologies are used to manage study data, track samples, and maintain quality control, including personal data assistants, barcodes, global information systems, and fingerprint scans. Close collaboration with the Nicaraguan Ministry of Health and use of almost entirely local staff are essential components for success. This study is providing critical data on the epidemiology and transmission of dengue in the Americas needed for future vaccine trials.

Basic Biostatistics for Geneticists and Epidemiologists: A Practical Approach: By Robert C. Elston and William D. Johnson

Hokanson, J. E. — 2009-06-24

RE: "CHANGES IN PERCEIVED JOB STRAIN AND THE RISK OF MAJOR DEPRESSION: RESULTS FROM A POPULATION-BASED LONGITUDINAL STUDY"

Smith, P. M., Beaton, D. E. — 2009-06-24

THREE AUTHORS REPLY

Wang, J., Schmitz, N., Dewa, C. — 2009-06-24

RE: "ABSTRACTS OF THE 42ND ANNUAL MEETING OF THE SOCIETY FOR EPIDEMIOLOGIC RESEARCH"

2009-06-24