CANCERS IN RELATIVES OF CHILDREN WITH NON-HODGKINS-LYMPHOMA

We undertook a family study of children treated at the Institute Gusta ve-Roussy in France to investigate a familial aggregation of cancer in the families of children with non-Hodgkin's lymphoma (NHL). We obtain ed family data for 284 children with NHL. Using the Standardized Incid ence Ratio, we compared the observed and expected number of families w ith at least one proband relative affected by cancer at a young age (b efore 46 years). We found a small but non-significant excess of all tu mors in first-degree relatives (SIR = 1.3, 95% Cl = 0.7-2.3) explained by a small but non-significant excess of hematological malignancies ( SIR = 1.5, 95% Cl = 0.2-5.5), particularly Hodgkin's disease and leuke mia, and of osteosarcoma (SIR = 7.5, 95% Cl = 0.1-41.4), This is proba bly a lower bound of the SIR, because the expected number of families was estimated from cancer incidence in France between 1978 and 1982, w hereas most cancers occurred before this period. Other tumors were not in excess in first-degree relatives.