Heritability of Hematologic Neoplasms in Twins: An Update

Abstract 3636

Hodgkin lymphoma, especially the young adult type, is one of the most heritable cancers. We previously reported a high risk of Hodgkin lymphoma to monozygotic (MZ), but not dizygotic (DZ) co-twins of cases, and only a modest difference in risk between MZ and DZ co-twins of non-Hodgkin lymphoma (NHL) cases (Mack, 1995). After an additional 18 years of follow-up, we have now updated the observed occurrence of hematologic malignancies in the initially unaffected co-twins of HL, NHL, multiple myeloma (MM), and chronic lymphocytic leukemia (CLL) twin probands. The number of calendar and age-specific person-years at risk for each co-twin was calculated from the date of diagnosis of the proband to the date of last follow-up of the co-twin, defined by the last date of contact, date of death ascertained directly or from linkage with the National Death Index, or evidence of current vital status from a national tracing program. The expected number of cases was calculated by applying the calendar and age-specific incidence rates by 5-year interval categories for each hematologic neoplasm from the Surveillance, Epidemiology and End-Results Program to the person-years of risk. Diagnoses of hematological neoplasms in the co-twins by age and year were ascertained by direct follow-up augmented by a linkage with the National Death Index, using diagnoses categorized by the ICD-9 codes. The standardized incidence ratio (SIR) was computed as the observed to expected number of cases; 95% confidence intervals (CI) (Breslow and Day, 1987), and the risk ratio (RR) (ratio of the SIR in MZ co-twins compared to that in DZ co-twins) were calculated. Whereas the SIR for DZ co-twins measures the heritable and acquired components of risk to first-degree family members, the RR provides evidence of genetic heritability, based on the additional genomic commonality of MZ twins.

A total of 364 (HL), 501 (NHL), 91 (MM), and 45 (CLL) co-twins of probands contributed to the analysis. The risk of developing the same hematologic neoplasm as the proband was generally higher in the MZ compared to the DZ co-twins, with the highest RR observed for HL (13.3) and the lowest for NHL (1.75). Although more than 10,000 person-years were added since the original paper, the RR's for HL and NHL did not change substantially from those reported in 1995. The RR for CLL was 3.3 suggesting moderately strong heritability. One MZ co-twin developed MM producing a large SIR, however chance cannot be easily ruled out. The findings of most interest are the continued very high risk of HL in MZ compared to DZ twins confirming the strong heritability of this neoplasm, and the relatively low RR for NHL. MZ and DZ co-twins of NHL probands had increased but similar SIR's, suggesting that shared environmental factors are more important than heritability. Subtype–specific information on the NHL type was not available from the ICD-9 codes used by the Death Index to identify new cases, so it is possible that stronger NHL heritability would be evident if subtypes were considered separately.