Familial risk of cancer shortly after diagnosis of the first familial tumor

Background: The diagnosis of the first cancer in a family may lead to the medical examination of the patient's relatives and the subsequent identification of additional familial cancers. If detection bias is present, familial risks could be overestimated soon after first diagnosis. Methods: We followed 1677722 offspring/siblings of 846448 probands from the year of diagnosis of the first familial tumor to the diagnosis of first cancer, death, emigration, or December 31, 2002, using the Swedish Family Cancer Database. The risks of cancer among the offspring and siblings of patients with melanoma and cancers of the breast, prostate, colorectum, cervix, and lung were compared with those in the general population. Relative risks (RRs) were determined using Poisson regression, according to the time after first diagnosis. All statistical tests were two-sided. Results: Daughters of women with breast cancer had a statistically significantly higher relative risk of in situ breast cancer during the year of the mother's diagnosis than they did 5 or more years later (RR = 4.78, 95% confidence interval [CI] = 2.16 to 10.6, 26.6 cases per 100 000, versus RR = 1.97, 95% CI = 1.65 to 2.37, 27.2 cases per 100 000; P = .033). Daughters diagnosed the same year as their mothers were younger and were diagnosed earlier in the calendar year than daughters of women diagnosed 5 or more years after their mothers. Similarly, the risk of invasive melanoma among the offspring of individuals with invasive melanoma was higher during the year of the parent's diagnosis than it was 5 or more years afterward (RR = 8.27, 95% CI = 3.82 to 17.9, 57.0 cases per 100 000, versus RR = 3.18, 95% CI = 2.55 to 3.97, 37.6 cases per 100000; P = .019). Sibling risks of in situ breast cancer, in situ cervical cancer, and invasive prostate cancer also decreased with time after diagnosis of the first familial tumor. Conclusions: Increased surveillance may result in the earlier detection of asymptomatic familial cancers, i.e., in detection bias. The possibility of overestimated familial risks of cancer shortly after diagnosis of the first familial tumor should be considered before a patient's clinical and genetic counseling is implemented.