Research projects

Family History Working group

Working Group Participants: Ann G. Schwartz, Neil Caporaso, Carol Etzel, Michele Cote

A. Background and Specific Aims

    First-degree family history of lung cancer is a well-established risk factor for lung cancer, increasing risk approximately 2-fold. Less well-established is the role of family history of lung cancer in minority populations and how much risk varies by age at diagnosis and in never smokers. In addition, there has been little consensus about what other cancers aggregate with lung cancer in families. The specific aims of the proposed study are to:

  1. Estimate the contribution of family history to risk of developing lung cancer in various population groups, by age of the proband and in never smokers.
  2. Estimate cumulative risk of cancer (lung and others) in family members of lung cancer cases by age, race/ethnicity, and smoking status.
B. Methods

    The collection of family history data by participants in ILCCO is quite variable. Eleven studies have information on 1) whether any first degree relative has lung cancer or any cancer (and include site of that cancer), and whether that relative is a parent, sibling or child. From the information provided, 9082 cases (1211 with a first-degree family history) and 9827 controls have this type of information. In this group, a traditional case-control study will be conducted. The variable of interest will be family history yes/no, defined in several ways including lung cancer, any cancer, and specific cancer types, and by number of relatives affected. Analyses will also be conducted after stratification by race/ethnicity, age of diagnosis, and histologic type.

    The analyses described, however, do not account for the age, sex, and smoking histories of relatives, all of which alter risk of lung cancer substantially. To estimate cumulative risk of lung or other cancers in family members of lung cancer cases, a subset of the ILCCO studies will be used. In at least three studies, the numbers of relatives, numbers of affected relatives and at least smoking status (ever/never) of the relatives are available. This subset of studies includes over 5000 cases and 5000 controls, including a large number of early onset and non-smoking cases, as well as minorities. In these studies, risk of cancer in relatives will be the outcome, with the variable of interest being whether the relative is related to a case or control. Analyses will estimate relatives' risk of lung cancer and other cancers adjusting for smoking in the relatives. Stratified analyses by race/ethnicity, age, and smoking status will also be conducted.

C. Variables Needed

    See attached spreadsheet. We are requesting data from all first-degree relatives (parents, offspring, full siblings), regardless of whether or not the relative has had cancer.