一种病例对照家系资料的家庭相关统计分析方法

目的 提出一种分析病例对照家系资料家庭相关的统计方法。方法 同时建立先证者疾病状态和给定先证者疾病状态条件下亲属疾病状态的边际均数回归模型和亲属的边际关联回归模型,联合条件OR值和边际OR值评价家庭相关。结果 该模型参数解释符合抽样设计的特点,在评价危险因素作用上能充分利用信息,因此有较高效能。此外,该法具有GEE2方法估计家庭相关的所有其他优点。结论 该法用于病例对照家系资料估计疾病的家庭相关,简单、方便、有效。     

应用边际回归模型研究原发性肝癌的家庭相关

目的探讨泰兴市原发性肝癌的家庭聚集性。方法采用病例对照家系设计收集家系资料，应用边际回归模型方法分析亲属间肝癌的家庭相关大小和模式。结果肝癌病例的亲属比对照亲属有更高的疾病风险。调整了乙肝表面抗原后，亲属间的相关不再有统计学意义。结论边际回归模型可用于家系资料，研究表明肝癌存在家庭聚集性，HBV感染在肝癌的家庭聚集性中起着很重要的作用。     

病例对照家系研究中发病年龄的家庭相关分析方法

目的 探讨失效时间的关联测量和病例对照家系研究中发病年龄家庭相关的分析方法。方法 分层Cox模型用于估计病例对照家系研究中肝癌发病年龄的交叉比。结果 显示先证者和父亲、母亲、同胞间发病年龄的家庭相关有统计学意义，而先证者和配偶间发病年龄的家庭相关没有统计学意义。结论 分层Cox模型可用于病例对照家系研究中发病年龄的家庭相关估计。     

脑梗塞家族聚集性的病因和流行病学研究

本研究通过遗传流行病学病例对照方法，对宜兴市２１４个家系（１０９个脑梗塞先证家系和１０５个对照家系）进行了脑梗塞家族史的遗传易感性研究，结果表明：病例和对照家系有良好的均衡可比性，病例组一级亲属总的患病率为５．１２％，与对照组的１．６３％相比，差异有非常显著性，不论其父母或同胞，均以病例亲属的患病率显著高于对照，而子女中两者无显著性差异；二项分布显示，脑梗塞家族中实际发病数超过其二项分布的理论概率     

食管癌的遗传流行病学研究

目的：探讨遗传因素与食管癌的关系。方法：通过遗传流行病学病例对照研究方法,对海安县２６５个食管癌先证者家系和２６５个对照家系进行分离比、遗传度估计。结果：食管癌先证者一级亲属的发病率（５．３８％）明显高于对照一级亲属的发病率（２．８１％）,食管癌遗传度为２４．２４％,分离比为０．１２１４,小于０．２５,属多基因遗传疾病。结论：海安县食管癌的病因中,遗传因素起重要作用。     

Ⅱ型糖尿病家族聚集性的流行病学研究

目的　探讨家族遗传因素在Ⅱ型糖尿病发生发展中的作用。方法　采用遗传流行病学病例对照研究方法，对常州市３５０ 个家系（１８４ 个Ⅱ型糖尿病先证家系和１６６ 个对照家系） 进行了Ⅱ型糖尿病的遗传流行病学研究。结果　病例组一级亲属总的患病率为３．４７ ％ ，与对照组的１ ．０３ ％ 相比，差异有非常显著性（χ２ ＝ １７ ．６６ ，Ｐ＜ ０ ．０１） ；无论其父母、同胞或子女，均以病例亲属的患病率显著高于对照； Ⅱ型糖尿病家族中实际发病数超过其二项分布的理论概率范围，即Ⅱ型糖尿病的分布呈明显的家族聚集性； 单因素和多因素ｌｏｇｉｓｔｉｃ 回归模型拟合也提示，家族史仍是Ⅱ型糖尿病的最主要危险因素。结论　遗传因素在Ⅱ型糖尿病发病中占有重要地位     

Ⅱ型糖尿病的家庭聚集性研究

目的　通过对Ⅱ型糖尿病家庭聚集性的分析,探讨遗传因素在糖尿病患者一级亲属成员发病中所起的作用及其相对危险度。方法　采用以人群为基础的遗传流行病学病例对照研究方法,对３６３ 例Ⅱ型糖尿病先证者家系及２９１ 例人群对照家系进行了调查。结果　先证者家系一级亲属糖尿病的患病率为３ ．９４ ％ ,对照组一级亲属为１ ．０９ ％ ,相对危险度为３ ．６２ ,各组血缘亲属的相对危险度均在３ ．０ 以上;糖尿病先证者诊断时年龄越轻,其一级亲属的糖尿病患病率和相对危险度越高,其家系内发生多例糖尿病病例的可能性越大。结论　Ⅱ型糖尿病具有明显的家庭聚集倾向,其一级亲属对糖尿病的遗传易感性较高,是糖尿病预防和控制的重点人群。     

脑梗塞遗传病因的流行病学系列研究(Ⅰ)

通过遗传流行病学病例对照方法，对宜兴市２１４个家系（１０９个脑梗塞先证家系和１０５个对照家系）进行分离比、多基因遗传方式和遗传度的分析。结果表明：在病例和对照家系具有良好均衡可比性的前提下，脑梗塞的分离比为０．０５２９，显著低于０．２５，说明其不符合单基因遗传病的特征；多基因遗传方式的分析表明，其一级亲属的理论发病率为６．６７８％，与实际发病率８．６１１％相比，无统计学显著性差异，符合多基因遗传的特征；遗传度的估算显示；脑梗塞先证者的一级亲属中，遗传度为５３％，其中女性亲属的遗传度达５７％，显著高于男性亲属的３９．４％，说明在脑梗塞病因中，遗传因素起着一定的作用，特别是对女性作用更明显     

脑梗塞家族聚集性的研究

目的研究脑梗塞是否存在家族聚集性.方法采用病例对照研究,对哈尔滨市95个脑梗塞先证家系和95个对照家系进行了脑梗塞家族史的遗传易感性的研究.结果病例和对照家系年龄及性别有良好的均衡,病例组一级亲属总的患病率为16.94%,与对照组的3.70%相比,有显著性差异(x2=81.32,P＜0.01),不论其父母或同胞,均以病例亲属的患病率显著高于对照,而子女两者无显著性差异;二项分布分析显示,脑梗塞家族中实际发病数超过其二项分布的理论概率范围,即脑梗塞的分布呈明显的家族聚集性;单因素和多因素Logistic回归模型拟合也提示,遗传因素在脑梗塞发病中有重要地位.结论脑梗塞存在家族聚集性.     

Ⅱ型糖尿病家族聚集性的流行病学研究

目的探讨遗传因素与Ⅱ型糖尿病之间的关系。方法通过遗传流行病学病例对照研究方法,对常州市350个家系（184个Ⅱ型糖尿病先证家系和166个对照家系）进行了Ⅱ型糖尿病的遗传流行病学研究。结果病例和对照家系有良好的均衡可比性,病例组一级亲属总的患病率为3．47％,与对照组的1．03％相比,差异有非常显著性（X2=17．66,P＜0．005）,不论其父母、同胞或子女,均以病例亲属的患病率显著高于对照,二项分布显示,Ⅱ型糖尿病家族中实际发病数超过其二项分布的理论概率范围,即Ⅱ型糖尿病的分布呈明显的家族聚集性,单因素和多因素Logistic回归模型拟合也提示,遗传因素在Ⅱ型糖尿病发病中占有重要地位。结论遗传因素与Ⅱ型糖尿病的发生有一定的关系。     

Heterogeneity of breast cancer risk in families of young breast cancer patients and controls.

Familial heterogeneity of breast cancer risk was assessed among 4,159 first-degree female relatives of 1,074 population-based, young breast cancer cases (aged 20-54 years) diagnosed from December 1, 1980 to December 31, 1982 and 4,120 first-degree female relatives of 998 age- and race-matched, population-based controls from the metropolitan Detroit, Michigan, area. The family risk index method used for analysis considers family size, age, and race differences among families in the assessment of family risk. Families of cases showed a higher risk of breast cancer than did families of controls, with case families 1.80 to 4.24 times more likely to be defined as high risk than control families; the magnitude of the risk differential was dependent on the definition of high risk. Within the case families only, familial heterogeneity of risk was suggested, with a small proportion of families (less than 5%) at lower risk of breast cancer than most case families. A number of reproductive risk factors, age, race, and histologic type of cancer for the proband, and several family characteristics were investigated to help characterize the case families at higher and lower risk.     

Design considerations for association studies of candidate genes in families

In genetic epidemiologic studies, investigators often use generalized linear models to evaluate the relationships between a disease trait and covariates, such as one or more candidate genes or an environmental exposure. Recently, attention has turned to study designs that mandate the inclusion of family members in addition to a proband. Standard models for analysis assume independent observations, which is unlikely to be true for family data, and the usual standard errors for the regression parameter estimates may be too large or too small, depending on the distribution of the covariates within and between families. The consequences of familial correlation on the study efficiency can be measured by a design effect that is equivalent to the relative information in a sample of unrelated individuals compared to a sample of families with the same number of individuals. We examine design effects for studies in association, and illustrate how the design effect is influenced by the intra-familial distribution of covariate values such as would be expected for a candidate gene. Typical design effects for a candidate gene range between 1.1 and 2.4, depending on the size of the family and the amount of unexplained familial correlation. These values correspond to a modest 10% increase in the required sample size up to more than doubling the requirements. Design effect values are useful in study design to compare the efficiency of studies that sample families versus independent individuals and to determine sample size requirements that account for familial correlation.     

Familial clustering of obesity and breast cancer

One of the most striking characteristics of breast cancer (BC) is a tendency to familial aggregation. In order to evaluate whether familial clustering of obesity could account, at least in part, for the familial aggregation of BC, we compared the adult body size of entire sets of first-degree relatives belonging to 60 families with two or more cases of BC (case families) and 120 BC-free families (control families). Case families included an index case recently admitted for primary BC who had a confirmed first-degree family history for the disease. Control families included one population-based healthy index control with no family history and age-matched (2:1) to index cases. Index cases and controls, recruited from a pool of participants in a large case-control study, completed a questionnaire covering their own body size history as well as that of each of their first-degree relatives (598 case and 1,128 control relatives) using a validated system of body silhouette drawings. The odds ratio (OR) for premenopausal familial BC associated with having one parent markedly obese compared to none was 0.17 (95% confidence interval [Cl] 0.04–0.65), while having both parents obese resulted in an OR of 0.25 (95% Cl 0.04–1.56). Obesity among siblings was not related to premenopausal familial BC risk nor was familial obesity a significant predictor of familial BC after menopause. Index cases from both menopausal groups tended to be thinner than their unaffected relatives at age 40 years and thereafter. The inverse relationship between parental obesity and premenopausal BC risk is concordant with the protective effect of obesity on early-onset BC previously reported at the individual level. © 1996 Wiley-Liss, Inc.     

Application of odds ratio regression models for assessing familial aggregation from case-control studies

A regression model for estimating covariate effects on odds ratios to test for familial aggregation of common disease in first-degree relatives of cases and controls is presented and illustrated by using family data from a study of chronic obstructive pulmonary disease. These estimators are in essence an extension of the Mantel-Haenszel estimator of odds ratio but do not require the assumption of independence among relatives. A robust test statistic for possible effects of covariates such as the matching variables for cases and controls on odds ratio is also presented. In data on 156 adult first-degree relatives of 28 cases with demonstrated airway obstruction and 28 controls matched for age, sex, race, and hospital status, there appeared to be a difference in the odds ratio among families of black and white case-control pairs. However, the small sample size available prevents conclusive interpretation of this observation.     

Some further results on incorporating risk factor information in assessing the dependence between paired failure times arising from case-control family studies: an application to prostate cancer

In a typical case-control family study, detailed risk factor information is often collected on cases and controls, but not on their relatives for reasons of cost and logistical difficulty in locating the relatives. The impact of missing risk factor information for relatives on estimation of the strength of dependence between the disease risk of pairs of relatives is largely unknown. In this paper, we extend our earlier work on estimating the dependence of ages at onset between paired relatives from case-control family data to include covariates on cases and controls, and possibly relatives. Using population-based case-control families as our basic data structure, we study the effect of missing covariates for relatives and/or cases and controls on the bias of certain dependence parameter estimators via a simulation study. Finally we illustrate various analyses using a case-control family study of early onset prostate cancer.     

Correlated binomial variates: properties of estimator of intraclass correlation and its effect on sample size calculation.

In group randomized studies, the sample size calculations are complicated by within group (worksite, community, etc.) correlation. We compare by simulation the moment method and the more standard ANOVA method of estimating the intraclass correlation. We find the former is less biased for a small to moderate number of clusters but the difference disappears when the appropriate degree of freedom is used for the ANOVA estimator. We propose a simulation approach for sample size determination and illustrate it with an example.     

Polygenic inheritance of breast cancer: Implications for design of association studies

Susceptibility to breast cancer is likely to be the result of susceptibility alleles in many different genes. In particular, one segregation analysis of breast cancer suggested that disease susceptibility in noncarriers of BRCA1/2 mutations may be explicable in terms of a polygenic model, with large numbers of susceptibility polymorphisms acting multiplicatively on risk. We considered the implications for such a model on the design of association studies to detect susceptibility polymorphisms, in particular the efficacy of utilizing cases with a family history of the disease, together with unrelated controls. Relative to a standard case-control association study with cases unselected for family history, the sample size required to detect a common disease susceptibility allele was typically reduced by more than twofold if cases with an affected first-degree relative were selected, and by more than fourfold if cases with two affected first-degree relatives were utilized. The relative efficiency obtained by using familial cases was greater for rarer alleles. Analysis of extended families indicated that the power was most dependent on the immediate (first-degree) family history. Bilateral cases may offer a similar gain in power to cases with two affected first-degree relatives. In contrast to the strong effect of family history, varying the ages at diagnosis of the cases across the range of 35-65 years did not strongly affect the power to detect association. These results indicate that association studies based on cases with a strong family history, identified for example through cancer genetics clinics, may be substantially more efficient than population-based studies.     

Familial factors associated with malignant gliomas

Family histories of male patients with histologically confirmed malignant gliomas were compared to family histories of controls (wives). Included were 77 case families with 892 relatives and 77 control families with 719 relatives. Cases had significantly more siblings than controls (P = 0.02), although cases were not preferentially the oldest or the youngest sibs. Odds ratios of two or more were found for mental retardation, Parkinson's disease, and meningitis for the relatives of cases versus controls, but none were statistically significant. The excesses of Parkinson's disease and meningitis were explained by the family of one particularly interesting case containing three relatives with meningitis and two relatives with Parkinson's disease. Noteworthy age-adjusted odds ratios for cancer among relatives of cases compared to relatives of controls were 1.6 (95% confidence interval (CI) = 1.0-2.3) for cancer of any site, 2.4 (95% CI = 0.8-6.1) for breast cancer, and 4.0 (95% CI = 0.6-10.7) for lung cancer. Only the odds ratio for cancer of any site was statistically significant. Overall, 6 of 77 (8%) of cases came from families that included two or more relatives with breast or lung cancer in addition to the proband with malignant glioma. These three cancer sites may form familial clusters worthy of further evaluation in future studies by pedigree and genetic linkage analyses.     

Comparison of family history measures used to identify high risk of coronary heart disease

We examined 15 published continuous family history measures (scores) as well as two new formulations in terms of several desirable properties. We applied the scores to sample pedigrees and found that some systematically increase with family size. In contrast to aggregate scores, non-aggregate scores are sensitive to the age, sex, and covariate status of individual relatives but are unstable when the families are small. We also applied these scores to our own population case-control data, characterised by a high proportion of missing and false-negative responses. In these small families, all scores provided significant discrimination between CHD cases and controls beyond the usual categorical definition of positive family history, but appeared no better than detailed categorical definitions or even simple counts. Our new formulations offer no solution to the problems of few data; most scores apply asymptotic approximations to differences between observed and expected number of affected relatives and are not suited to small families. All scores would be improved by ruling out families with only one affected relative, as is being done in the NHLBI Family Heart Study. We recommend that researchers, when using a family history measure, consider the number of informative families and other characteristics of their data prior to choosing any particular formulation.