To determine whether the occurrence of first and second primary malignancies in first-degree relatives of cancer patients aged 0 -25 years (probands) differed from that in the general population, a cohort study was carried out on 860 relatives of 265 probands living in the province of Trieste, Italy. During the follow-up period (median duration ؍ 28 years, 25th-75th percentile ؍ 20 -34), the relatives developed 103 first primary cancers vs. 88.9 expected for a standardized incidence ratio ( The causes of most malignant tumors of childhood remain little understood. Few environmental factors are consistently associated with cancers affecting children, and Mendelian disorders play a role in the etiology of only a small proportion of them. 1 Epidemiologic studies of familial aggregation are carried out to determine whether clustering is real or due to chance association, and constitute the first of the 3 sequential steps 2 aimed at clarifying genetic-environmental interaction in disease etiology. They have failed to provide strong evidence that the most frequent childhood malignancies cluster in families and have not reported an overall increased risk of cancer among the close relatives of children with tumors. 1 However, we observed a statistically significant excess of first primary malignancies in the first-degree relatives of childhood cancer patients and, in particular, among the relatives of children affected by brain tumors, through a population-based cohort study conducted in the northeastern Italian province of Trieste. 3 The data presented here update those in Brunetti et al. 3 Furthermore, we estimate the likelihood of developing a second primary neoplasm in those relatives, and the risk of first and second primaries among the first-degree relatives of the individuals who had cancer between 15 and 25 years of age, because tumors in adolescents and young adults were found to have several epidemiologic features in common with those occurring in childhood. 4
SUBJECTS AND METHODSAll resident males and females aged 0 -25 years with a first primary cancer diagnosed between 1 January 1971 and 31 December 1993 (probands) were sorted out from the file of the Cancer Registry of the Province of Trieste (CRPT). 5 Their tumors were grouped according to the International Classification of Childhood Cancer, 6 a histology-based scheme that also proved suitable for the descriptive analysis of neoplasms among adolescents and young adults. 4 No cancer was diagnosed incidentally, registered on the basis of information on the death certificate only (DCO) or recorded with an unknown or ill-defined primary site. 7 The proportion of tumors with microscopic verification of diagnosis was 95%, because 1 glioma of the optic chiasma, 1 supratentorial brain tumor, 2 craniopharyngiomas, 3 brainstem neoplasms and 7 pituitary microadenomas were detected by means of imaging techniques. At the date of study termination (31 December 1999) 1 proband (0.4%) was lost to follow-up because she had emigrated abroad.Information found in medic...