Family history of ovarian cancer patients: a case-control study.

In order to obtain a more correct estimate of cancer incidence in relatives of cases with ovarian cancer and relatives of age-matched controls, information was obtained on family size, date of birth and death, and cancer history. The average total number of first degree (parents and siblings) and some second degree (uncles and aunts) relatives for cases (15.03) and controls (15.22) is very similar. Of all the relatives 33% were Alberta residents or decreased in Alberta after 1966 so they could be checked in the population based Alberta Cancer Registry. There was a 14% level of error in the information provided, 9% for first degree relatives and 25% for uncles and aunts. The errors included missed malignancies, benign lesions quoted as malignancies, incorrect cancer site and inaccurate dates of birth, diagnosis or death. Results were compared between the relatives of cases and controls using either all relatives or only Alberta residents. For the Alberta residents, expected numbers of malignancies were calculated by age/sex-specific person/years and age/sex-specific incidence rates. The number of relatives with ovarian cancer is significantly higher in cases than in controls in both data sets, but shows a limited relative risk of 2.61 (95% confidence limits 1.12-1.59). Clustering of cases is exceptional, only one patient out of 197 had two relatives with ovarian cancer and this was a case with a larger than average total number of relatives. In conclusion, although there is a significantly higher incidence of ovarian cancer in the family history of cases than of controls, it is not very common and only nine out of 197 cases (4.6%) presented with such background.