Familial risks for cancer with reference to lung cancer

Sammanfattning: Familial aggregation of cancer may be due either to environmental factors shared by family members, or to shared genes. Familial clustering has been an avenue to the understanding of the etiology of cancer and has been a basis for clinical decisions and counseling, as well as guiding the identification of cancer-related genes. The nation-wide Swedish Family-Cancer Database was created by linking the Statistics Sweden Multi- Generation Register with the Swedish Cancer Registry, national census data and death notifications. All Swedes born after 1931 and their parents are included in the Database, which encompasses over 10 million individuals. Cancer cases were retrieved from the Swedish Cancer Registry from the years 1958, including over 1 million primary cancers and in situ tumors (paper I). The Family-Cancer Database was used as the source of family and cancer data, according to the premise that comparisons can be done based on different probands: parental probands are informative of dominant effects in offspring, whereas analysis of risks between siblings who lack affected parents provides clues about possible recessive effects. With specific reference to lung cancer, in paper II, we used the Family-Cancer Database to examine the time trends of lung cancer in Sweden by histological type, with specific reference to gender, birth cohort, years of diagnosis (period), and age. The incidence rates of squamous cell carcinoma and other subtypes in men peaked in the period 1980-1990, and then decreased slightly thereafter. In contrast, the incidence rates of adenocarcinoma have continued to increase throughout the period. In women, all subtypes except squamous cell carcinoma have been increasing linearly. The incidence ratio of male to female for all cases of lung cancer was 2.8. For squamous cell carcinoma, the ratio was 12.4 in the beginning and 3.6 at the end. For adenocarcinoma, the ratio was stable at about 1.5 for the whole period. In paper III, the standardized incidence ratio was 1.87 (95% CI 1.66-2.10) for all offspring (0 66 years old) when parents had lung cancer. The proportion of familial affect was 6.39%, and the population-attributable risk of all lung cancers was 2.97%. Lung cancer in offspring was associated with parental rectal, cervical, kidney, urinary bladder and endocrine gland tumors. Age-specific, histology-specific familial risks for lung cancer were analyzed in paper IV. For offspring diagnosed before 50 years, the SIRs for histological types of lung cancer between offspring and parents were 1.98 for adenocarcinoma (offspring) and small cell/large cell carcinoma (parents), 2.66 for squamous cell carcinoma and small/large cell carcinoma (offspring/parents), 3.54 for small cell carcinoma and squamous cell carcinoma, and 3.70 for large cell carcinoma and adenocarcinoma. At a young age, risks between siblings were higher than those between offspring and parents. The SIR ratio (sibling risk/offspring risk) was 2.92 for all lung cancers. The early onset familial component accounted for 29.4% of familial adenocarcinoma and 33.3% of familial small and large cell carcinoma. The proportion was lowest for squamous cell carcinoma (13.3%). In papers III and V, for multiple primary lung cancers we also found that risk of lung cancer was increased after upper aerodigestive tract, breast, cervical, kidney, urinary bladder or squamous cell skin cancer, or non- Hodgkin s lymphoma, Hodgkin s disease or leukemia, through all follow-up periods in men and women. Patients had an increased familial risk when their first-degree relatives were diagnosed with two lung cancers compared to those whose first-degree relatives were diagnosed with one lung cancer, and vice versa. In summary, we suggested that there is an equal sensitivity of both genders to tobacco-induced lung cancer. Familial history of lung cancer was associated with increased risk of lung cancer; the population-attributable fraction of familial lung cancer was 2.97%. A large proportion of lung cancers before 50 years of age appear to be heritable and are probably due to a high-penetrant recessive gene or genes that predispose to tobacco carcinogens. Familial risks for multiple primary lung cancers have also suggested that there is an inherited susceptibility.