Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Reproductive History”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 649 records · Page 36Linked to original sources
PubMed2005

Place of preimplantation diagnosis in genetic practice.

Preimplantation genetic diagnosis (PGD) is currently one of the practical options available for couples at-risk to avoid the birth of children with genetic and chromosomal disorders. Despite its novelty, PGD has already become an alternative to traditional prenatal diagnosis, allowing establishing only unaffected pregnancies avoiding the risk for pregnancy termination. Indications for PGD have currently expanded beyond those practices in prenatal diagnosis, such as late-onset diseases with genetic predisposition, and preimplantation HLA typing with the purpose of establishing potential donor progeny for stem cell treatment of siblings, which makes PGD also an important compliment to prenatal diagnosis. The fact that more than 1,000 apparently healthy unaffected children have been born after PGD suggests its accuracy, reliability, and safety. PGD is presently an excellent option for carriers of balanced translocations, and appears to be of special value for avoiding age-related aneuploidies in patients of advanced reproductive age. The accumulated experience of thousands of PGD cycles for poor prognosis in vitro fertilization (IVF) patients provides strong evidence of the improvement of clinical outcome, particularly obvious from the reproductive history of patients. This makes of practical relevance to inform couples at-risk about availability of PGD option, so they make their own choice in avoiding the birth of affected offspring and having healthy children of their own.

Chromosome Disorders↗
PubMed1981

Patterns of growth and malnutrition among preschoolers in Belize.

An anthropometric survey of 750 preschoolers was conducted 1979 in Belize to assess nutritional status in two districts, one coastal and the other inland. Mothers were interviewed concerning reproductive history, child's health and dietary history, and associated factors. Analysis included the recognition of low arm circumference, weight-for-age, stature-for-age, and weight-for-stature. Survey results indicate that about 25% of the children, birth to 5 years old, show evidence of stunting, while 2.5% show evidence of wasting. Comparisons of rates of malnutrition by ethnic group reveal the Maya and Garifuna (Black Carib) children are significantly more commonly malnourished than others. Applying a discriminant function to groups of children designated as "poor" and "better-than-average" growth classes reveals several factors which may be contributory to the etiology of preschool malnutrition in Belize. the frequency of diarrhea and age at introduction to sold foods, are significantly related to growth retardation among preschoolers. Analysis of 24-hour-diet recalls corroborates the patterning of malnutrition among the ethnic groups.

Age Factors↗
PubMed1983

Airport noise exposure and the postnatal growth of children.

To determine whether chronic exposure to airport noise affects children, a study was conducted of the physical growth of children, aged 5-13 years, from two communities, one exposed to airport noise (n = 148) and another, not exposed (n = 102). Ten standard anthropometric measurements were made according to U.S. Health Examination Survey guidelines, and information on the social and biological characteristics of each family was collected in interviews. Hotelling's T2-tests were performed comparing the noise-exposed and nonnoise-exposed samples. There was no significant difference between the sample for measurements of social and biological characteristics of the families (including socioeconomic status and maternal reproductive history). Parental anthropometrics differed significantly (T2 = 24.32, P = 0.0001) as did child anthropometrics (T2 = 21.01, P = 0.032). For the child anthropometrics, noise-exposed children's slightly smaller measures of body bulk, together with their larger facial breadths, contributed to the significant T2. When the entire sample (n = 250) was analyzed by multiple linear regression, noise exposure was a significant predictor only of male triceps and subscapular skinfolds. For these two variables, and most other anthropometrics, however, the beta coefficients were negative for both sexes. In order to include information on the covariance structure among all variables, a canonical correlation analysis was performed. Noise exposure loaded negatively on the third canonical variate and was paired with positively loaded measures of body bulk. Taken together, the three analyses suggest that while there is some evidence for a slight effect of airport noise on measures of body bulk for males in this sample, most measures of postnatal growth for both males and females were unaffected by the noise levels experienced.

Adolescent↗
PubMed1985

Variation in size at birth and cigarette smoking during pregnancy.

The influence of cigarette smoking during pregnancy and other familial factors on size at birth and gestation length is investigated among 458 births to 227 mothers living in a suburban community in the U.S. In this sample, 56% of the births were to mothers who reported not smoking during the pregnancy and 35% were to mothers who reported smoking 20 cigarettes or less. Multiple stepwise regression analysis was employed to examine the influence of cigarette smoking after statistical adjustment for such social and biological characteristics of the family as parents' sizes, education, income, and aspects of mother's reproductive history. After correction for significant social and biological characteristics, smoking status was a significant contributor to birth weight variation. In fact, cigarette smoking had the next-largest partial correlation coefficient (r = -0.26) second to gestation length. Birth length is also negatively associated with cigarette smoking, though not so strongly as is birth weight. The reduction in birth lengths can be attributed to the reduction in birth weights. Gestation length was not associated with cigarette smoking in this sample. The analysis of collinearity between smoking status and the other independent variables indicates that the effect of smoking appears to be independent of interrelationships among the independent variables.

Birth Weight↗
PubMed1990

Sources of variation in interbirth intervals among captive bonnet macaques (Macaca radiata).

Female fitness is a function of variation in the length of females' reproductive careers, the viability of their offspring, and the frequency with which they give birth. Infant loss shortens interbirth intervals in most primate species, but we know considerably less about other factors that contribute to variation in the length of interbirth intervals within groups. In one large captive group of bonnet macaques, maternal parity, age, experience, family size, and recent reproductive history are all associated with variation in the length of intervals that follow the birth of surviving infants. Primiparous females have the longest interbirth intervals, while multiparous females who have produced surviving infants in the past and have raised their last infant successfully have the shortest interbirth intervals. Infant sex and maternal rank have no direct effect upon the length of interbirth intervals. One of the underlying causes of variation in the length of interbirth intervals after surviving births seems to be variation in the timing of conceptions among females. Females who conceive early in the mating season tend to have shorter interbirth intervals than other females. However, females who are multiparous, experienced, and have recently raised infants have late conceptions and short interbirth intervals.

Age Factors↗
PubMed1994

Neonatal size and infant mortality at high altitude in the western Himalaya.

A prospective study was undertaken in Ladakh, India, a high-altitude region of the Himalaya, to investigate the effects of small average birth size on neonatal mortality. While such studies exist from high-altitude regions of the New World and shed light on the adaptive status of high-altitude-dwelling populations there, this is the first to examine this relationship in the Himalaya. In a sample of 168 newborns, birthweight and other anthropometric measurements were reduced relative to Andean and Tibetan newborns. Logistic regression and hazard analysis showed that neonatal biological characteristics such as weight, fatness, and circumferences were important predictors of survival probabilities of infants, especially in the neonatal period. Low Rohrer's Ponderal Index (PI) was particularly strongly related to poor survival outcome. Males and females showed no significant differences in mortality risk. Data derived from reproductive histories revealed that neonatal mortality accounted for 70-80% of total infant mortality in Ladakh. Compared to other high-altitude studies, small newborn size in Ladakh was associated with much higher mortality risks; mortality risk rose dramatically with birthweights below the mean (2,764 grams), which characterized 50% of all newborns. It is argued that newborns in Ladakh are subject to strong directional selective forces that favor higher birthweights that incur lower risks of neonatal mortality, while Andean infants are subject to relatively mild selection pressure at both ends of the birthweight distribution. Given the overall small size at birth of Ladakhi newborns and the poor survival outcomes of newborns below the mean, it is suggested that this population is less well adapted in a biological sense to the stresses inherent in this high-altitude environment than are Andean populations, perhaps due to the relatively recent colonization of the area and the substantial genetic admixture that has occurred in the past.

Altitude↗
PubMed2005

Breast cancer in sub-Saharan Africa: how does it relate to breast cancer in African-American women?

BACKGROUND: African-American women have had a lower incidence, yet higher mortality rate from breast cancer compared with White-American women. African-American women also have had a higher risk for early-onset, high-grade, node-positive, and hormone receptor-negative disease. Similar features have characterized hereditary breast cancer, prompting speculation that risk factors could be genetically transmitted. Further evaluation of this theory required the study of breast cancer among women from sub-Saharan Africa because of their shared ancestry with African-American women. METHODS: Publications from 1988 to 2004 of English-language literature on breast cancer in Africa were reviewed. RESULTS: Women from sub-Saharan Africa were found to have a low incidence of breast cancer. This was partly explained by a largely protective reproductive history, including late menarche, early menopause, high parity with prolonged breastfeeding, irregular menses, and fewer ovulatory cycles. The average age at diagnosis, however, was approximately 10 years younger than breast cancer patients of western nations, and disease stage distribution was shifted toward more advanced disease, which resulted in higher mortality rates. These features were found to be similar to data on breast cancer in African-American women. Mutations in BRCA1 and BRCA2 have been reported in African-American women, but the extent of the contribution of BRCA1 and BRCA2 to breast cancer burden in Africa was uncertain. Limited financial resources lead to suboptimal cancer data collection, as well as delayed diagnosis and treatment of many African breast cancer patients. CONCLUSIONS: Parallels between breast cancer burdens of African-American and sub-Saharan-African women were provocative, indicating the need for further exploration of possible genetically transmitted features related to estrogen metabolism and/or breast cancer risk.

Africa South of the Sahara↗
PubMed1984

Complete hydatidiform mole in Hawaii: an epidemiological study.

An analysis of hydatidiform moles occurring in Hawaii over a 14-year period (1968-1981) was undertaken. The pathology of all 278 reported molar pregnancies was reviewed and showed 69.4% to be complete, 24.5% to be partial, and 6.1% to be nonmolar. A case-control epidemiological investigation of the complete moles showed maternal age and race to be important, moles being significantly more prevalent in women under 20 and over 40 yr of age and also in women of Japanese, Filipino, and other Oriental ancestry. However no difference was seen in the prevalence of moles between Oriental women born in the Orient and those born and raised in Hawaii. No significant difference was found in paternal age, paternal race, socioeconomic status, or reproductive history, suggesting that these factors do not play an important role in the etiology of complete hydatidiform mole. Incidence rates for complete moles were calculated taking age and race into consideration and ranged from a high of 1 in 150 to a low of 1 in 2,000 naturally terminating pregnancies.

Adolescent↗
PubMed2004

Increased risk of breast cancer following different regimens of hormone replacement therapy frequently used in Europe.

Epidemiologic studies have shown an increased risk of breast cancer following hormone replacement therapy (HRT). The aim of this study was to investigate whether different treatment regimens or the androgenecity of progestins influence the risk of breast cancer differently. The Danish Nurse Cohort was established in 1993, where all female nurses aged 45 years and above received a mailed questionnaire (n = 23,178). A total of 19,898 women returned the questionnaire (86%). The questionnaire included information on HRT types and regimens, reproductive history and lifestyle-related factors. Breast cancer cases were ascertained using nationwide registries. The follow-up ended on 31 December 1999. Women with former cancer diagnoses, women with missing information on HRT, surgical menopause, premenopausal, as well as hysterectomized women were excluded, leaving 10,874 for analyses. Statistical analyses were performed using Cox proportional hazards model. A total of 244 women developed breast cancer during follow-up. After adjustment for confounding factors, an increased risk of breast cancer was found for the current use of estrogen only (RR = 1.96; 95% CI = 1.16-3.35), for the combined use of estrogen and progestin (RR = 2.70; 95% CI = 1.96-3.73) and for current users of tibolone (RR = 4.27; 95% CI = 1.74-10.51) compared to the never use of HRT. In current users of combined HRT with testosterone-like progestins, the continuous combined regimens were associated with a statistically significant higher risk of breast cancer than the cyclical combined regimens (RR = 4.16, 95% CI = 2.56-6.75, and RR = 1.94, 95% CI = 1.26-3.00, respectively). An increased risk of breast cancer was noted with longer durations of use for the continuous combined regimens (p for trend = 0.048). The European traditional HRT regimens were associated with an increased risk of breast cancer. The highest risk was found for the use of continuous combined estrogen and progestin.

Aged↗
PubMed2004

Smoking is associated with postmenopausal breast cancer in women with high levels of estrogens.

We investigated the association between smoking and risk of postmenopausal breast cancer in groups defined by high levels of estrogens, a factor known to enhance tumour progression. Two prospective cohorts of Swedish women provided 260 postmenopausal breast cancer cases and 514 controls. Blood samples were collected at baseline, and anthropometry, life-style factors and reproductive history had been assessed. Subjects were classified into quartiles with regard to the level of estrone, and into three categories with regard to estradiol. All analyses of the relation between smoking and breast cancer were repeated in different categories of these hormones. Logistic regression analysis, adjusted for matching factors, i.e., age at baseline, storage time and sub-cohort, yielded odds ratios (OR) with 95% confidence intervals (CI). Ever-smoking was associated with breast cancer in the top category of estrone, 2.02 (1.17-3.49). The highest risk was seen among ex-smokers, 2.96 (1.53-5.75). The pattern was similar for estradiol. Recent smoking cessation was associated with a high OR in top categories of estrone, 4.38 (1.27-15.2) and estradiol 10.0 (1.14-88.7). Smoking initiation before the age of 20 was associated with breast cancer in the top category of estrone, 2.73 (1.27-5.91). Several potential confounders were introduced into the statistical model, but none remained using backward selection. We conclude that ever-smoking was associated with the risk of breast cancer in women with high levels of estrone, and that ex-smoking was associated with breast cancer in women with high levels of estrone or estradiol.

Aged↗
PubMed2005

Prospective study of alcohol consumption and breast cancer risk in Japanese women.

Epidemiologic evidence is lacking for the association between alcohol consumption and the risk of breast cancer in Japanese women. We addressed this association in a prospective cohort study with an average follow-up of 7.6 years. At baseline (1988-1990), cohort participants completed a self-administered questionnaire that included alcohol use, reproductive history and hormone use. The women were followed up for breast cancer incidence through December 31, 1997. Cox proportional hazards models were used to calculate relative risks (RRs) and 95% confidence intervals (CIs) for breast cancer incidence and any association with alcohol consumption. During a follow-up of 271,412 person-years, we identified 151 women with breast cancer, of whom 45 were current drinkers and 11 drank > or =15 g of alcohol/day. After adjustment for age and other potential risk factors for breast cancer, the RR for current drinkers was 1.27 (95% CI 0.87-1.84) compared to nondrinkers. Average alcohol intake of <15 g/day did not significantly increase the risk for breast cancer. However, risk was significantly increased for women who consumed > or =15 g/day of alcohol (RR = 2.93, 95% CI 1.55-5.54). Age at starting drinking and frequency of consumption per week were not significantly associated with breast cancer risk. Our cohort study demonstrated that Japanese women who consume at least a moderate amount of alcohol have an increased risk of breast cancer.

Adult↗
PubMed2005

Effect of pregnancy as a risk factor for breast cancer in BRCA1/BRCA2 mutation carriers.

Early age at first birth and multiparity have been associated with a decrease in the risk of breast cancer in women in the general population. We examined whether this relationship is also present in women at high risk of breast cancer due to the presence of a mutation in either of the 2 breast cancer susceptibility genes, BRCA1 or BRCA2. We performed a matched case-control study of 1,260 pairs of women with known BRCA1 or BRCA2 mutations, recruited from North America, Europe and Israel. Women who had been diagnosed with breast cancer were matched with unaffected control subjects for year of birth, country of residence, and mutation (BRCA1 or BRCA2). Study subjects completed a questionnaire detailing their reproductive histories. Odds ratios (ORs) and 95% confidence intervals (CIs) were derived by conditional logistic regression. Among BRCA1 carriers, parity per se was not associated with the risk of breast cancer (OR for parous vs. nulliparous = 0.94; 95% CI = 0.75-1.19; p = 0.62). However, women with a BRCA1 mutation and 4 or more children had a 38% decrease in breast cancer risk compared to nulliparous women (OR = 0.62; 95% CI = 0.41-0.94). In contrast, among BRCA2 carriers, increasing parity was associated with an increased risk of breast cancer; women with 2 or more children were at approximately 1.5 times the risk of breast cancer as nulliparous women (OR = 1.53; 95% CI = 1.01-2.32; p = 0.05). Among women with BRCA2 mutations and who were younger than age 50, the (adjusted) risk of breast cancer increased by 17% with each additional birth (OR = 1.17; 95% CI = 1.01-1.36; p = 0.03). There was no significant increase in the risk of breast cancer among BRCA2 carriers older than 50 (OR for each additional birth = 0.97; 95% CI = 0.58-1.53; p = 0.92). In the 2-year period following a birth, the risk of breast cancer in a BRCA2 carrier was increased by 70% compared to nulliparous controls (OR = 1.70; 95% CI = 0.97-3.0). There was a much smaller increase in breast cancer risk among BRCA2 carriers whose last birth was 5 or more years in the past (OR = 1.24; 95% CI = 0.79-1.95). A modest reduction in risk of breast cancer was observed among BRCA1 carriers with 4 or more births. Among BRCA2 carriers, increasing parity was associated with a significant increase in the risk of breast cancer before age 50 and this increase was greatest in the 2-year period following a pregnancy.

Adolescent↗
PubMed2006

Cancer risk among Danish women with cosmetic breast implants.

The available epidemiologic evidence does not support a carcinogenic effect of silicone breast implants on breast or other cancers. Data on cancer risk other than breast cancer are limited and few studies have assessed cancer risk beyond 10-15 years after breast implantation. We extended follow-up of our earlier cohort study of Danish women with cosmetic breast implants by 7 years, yielding 30 years of follow-up for women with longest implant duration. The study population consisted of women who underwent cosmetic breast implant surgery at private clinics of plastic surgery (n = 1,653) or public hospitals (n = 1,110), and a control group of women who attended private clinics for other plastic surgery (n = 1,736), between 1973-95. Cancer incidence through 2002 was ascertained using the Danish Cancer Registry. Risk evaluation was based on computation of standardized incidence ratios (SIR) and Cox proportional hazards models, adjusting for age, calendar period and reproductive history. We observed 163 cancers among women with breast implants compared to 136.7 expected based on general population rates (SIR = 1.2; 95% confidence interval [CI] = 1.0-1.4), during a mean follow-up period of 14.4 years (range = 0-30 years). Women with breast implants experienced a reduced risk of breast cancer (SIR = 0.7; 95% CI = 0.5-1.0), and an increased risk of non-melanoma skin cancer (SIR = 2.1; 95% CI = 1.5-2.7). Stratification by age at implantation, calendar year at implantation and time since implantation showed no clear trends, however, the statistical precision was limited in these analyses. When excluding non-melanoma skin cancer, the SIR for cancer overall was 1.0 (95% CI = 0.8-1.2). With respect to other site-specific cancers, no significantly increased or decreased SIR were observed. Similar results were found when directly comparing women who had implants at private clinics with women who attended private clinics for other plastic surgery, with rate ratios for cancer overall, breast cancer and non-melanoma skin cancer of 1.1 (95% CI = 0.8-1.6), 0.7 (95% CI = 0.4-1.3) and 1.5 (95% CI = 0.8-2.7), respectively. In conclusion, our study lends further support to the accumulating evidence that silicone breast implants are not carcinogenic. Reasons for the consistently reported deficit of breast cancer among women with breast implants remain unclear, whereas increased exposure to sunlight may explain the excess occurrence of non-melanoma skin cancer. We found no indication of delayed diagnosis of breast cancer due to the presence of breast implants.

Adolescent↗
PubMed1995

Multiple births and risk of breast cancer.

Data from a large, multicenter, population-based case-control study were analyzed to investigate the relation between multiple birth pregnancies and subsequent breast-cancer risk in the gravidas. Women less than 75 years old who had breast cancer were identified from statewide tumor registries in Wisconsin, western Massachusetts, Maine and New Hampshire. Controls aged less than 65 years were selected randomly from lists of licensed drivers, and controls aged between 65 and 74 were selected randomly from lists of Medicare beneficiaries. Information on reproductive history and other factors was obtained by means of telephone interviews. After excluding nulliparous women, 5,880 case subjects and 8,217 controls remained for analysis. Multiple birth pregnancies occurred in 146 cases and 218 controls. Adjusted for the number of full-term pregnancies and other confounders, the odds ratio (OR) for any multiple birth pregnancy was 0.94 (95% confidence interval, 0.75 to 1.17). The null association between multiple birth pregnancies and breast cancer remained if the mother's first pregnancy or last pregnancy resulted in a multiple birth. In addition, no trend in risk emerged for age at first multiple birth or time since last multiple birth. While other investigators have suggested biological mechanisms to explain both protective and detrimental effects of multiple births and subsequent development of breast cancer, this study does not support either assertion.

Adult↗
PubMed1995

Risk modifiers in carriers of BRCA1 mutations.

The majority of, but not all, women with mutations in the BRCA1 gene will be affected with breast or ovarian cancer by the age of 70. To establish whether known risk factors modify susceptibility to cancer in these women, we have studied the reproductive histories of 333 North American women who were found by haplotype analysis to carry BRCA1 mutations. An increased risk for breast cancer was associated with low parity and with recent birth cohort. The risk of ovarian cancer decreased with increasing age at last childbirth; however, in contrast to the case for sporadic cancer, the risk of ovarian cancer in BRCA1 carriers was found to increase significantly with increasing parity.

Adult↗
PubMed1995

Risk factors for hip fracture in European women: the MEDOS Study. Mediterranean Osteoporosis Study.

The aims of this study were to determine common international risk factors for hip fracture in women aged 50 years or more. We studied women aged 50 years or more who sustained a hip fracture in 14 centers from Portugal, Spain, France, Italy, Greece, and Turkey over a 1-year period. Women aged 50 years or more selected from the neighborhood or population registers served as controls. Cases and controls were interviewed using a structured questionnaire on work, physical activity, exposure to sunlight, reproductive, history and gynecologic status, height, weight, mental score, and consumption of tobacco, alcohol, calcium, coffee, and tea. Significant risk factors identified by univariate analysis included low body mass index (BMI), short fertile period, low physical activity. lack of sunlight exposure, low milk consumption, no consumption of tea, and a poor mental score. No significant adverse effects of coffee or smoking were observed. Moderate intake of spirits was a protective factor in young adulthood, but otherwise no significant effect of alcohol intake was observed. For some risks, a threshold effect was observed. A low BMI and milk consumption were significant risks only in the lowest 50% and 10% of the population, respectively. A late menarche, poor mental score, low BMI and physical activity, low exposure to sunlight, and a low consumption of calcium and tea remained independent risk factors after multivariate analysis, accounting for 70% of hip fractures. Excluding mental score and age at menarche (not potentially reversible), the attributable risk was 56%. Thus, about half of the hip fractures could be explained on the basis of the potentially reversible risk factors sought. In contrast, the use of risk factors to "predict" hip fractures had moderate sensitivity and specificity. We conclude that variations in lifestyle factors are associated with significant differences in the risk of hip fracture, account for a large component of the total risk, and may be of some value in selecting individuals at high risk.

Aged↗
PubMed1995

Clinical aspects of risk in women with endometrial carcinoma.

Carcinoma of the endometrium is the most common gynecologic malignancy, expected to account for 33,000 new cases and 6,000 deaths in 1995. Most endometrial cancers occur in postmenopausal women and produce abnormal vaginal bleeding. Some women exhibit the premalignant changes of atypical endometrial hyperplasia before developing an overt carcinoma. Identified epidemiologic risk factors include obesity, diabetes mellitus, use of unopposed exogenous estrogens, estrogen-secreting tumors, and a reproductive history characterized by prolonged estrogenic predominance. Diagnosis can be readily established by outpatient endometrial biopsy. Because clinical estimates of disease extent and spread are subject to substantial error, endometrial cancer is now a surgically staged neoplasm. A well-defined set of surgicopathologic risk factors have been incorporated into the staging scheme. Women with extrauterine disease comprise about 20% of cases and are at greatest risk for tumor recurrence and death from disease. Within the much larger group of women whose tumors are limited to the uterus, recurrence risk can be stratified by cytologic grade, cell type, depth of myometrial invasion, and extension to the cervix. About two-thirds of women have low-risk disease confined to the uterus when these criteria are employed, while the remaining one-third have high-risk subtypes. Recent areas of investigation have focused on molecular and genetic markers. Two clinical observations currently being examined are the poorer survival of Black women with uterine cancer and the apparent association of endometrial lesions with chronic tamoxifen suppression in women with breast carcinomas.

Biopsy↗
PubMed2001

Two-stage methods for the analysis of pooled data.

Epidemiologic studies of disease often produce inconclusive or contradictory results due to small sample sizes or regional variations in the disease incidence or the exposures. To clarify these issues, researchers occasionally pool and reanalyse original data from several large studies. In this paper we explore the use of a two-stage random-effects model for analysing pooled case-control studies and undertake a thorough examination of bias in the pooled estimator under various conditions. The two-stage model analyses each study using the model appropriate to the design with study-specific confounders, and combines the individual study-specific adjusted log-odds ratios using a linear mixed-effects model; it is computationally simple and can incorporate study-level covariates and random effects. Simulations indicate that when the individual studies are large, two-stage methods produce nearly unbiased exposure estimates and standard errors of the exposure estimates from a generalized linear mixed model. By contrast, joint fixed-effects logistic regression produces attenuated exposure estimates and underestimates the standard error when heterogeneity is present. While bias in the pooled regression coefficient increases with interstudy heterogeneity for both models, it is much smaller using the two-stage model. In pooled analyses, where covariates may not be uniformly defined and coded across studies, and occasionally not measured in all studies, a joint model is often not feasible. The two-stage method is shown to be a simple, valid and practical method for the analysis of pooled binary data. The results are applied to a study of reproductive history and cutaneous melanoma risk in women using data from ten large case-control studies.

Adolescent↗