No evidence for anticipation in lymphoproliferative tumors in population-based samples

Abstract

Genetic anticipation in familial non-Hodgkin's lymphoma, Hodgkin's lymphoma, and chronic lymphocytic leukemia (CLL) has been consistently reported in the literature. However, most of these findings were based on data from families ascertained for genetic studies. Fecundity bias, right censoring bias, and secular trends can lead to erroneous conclusions regarding the presence of anticipation. Our report investigates anticipation in four lymphoproliferative cancers, non-Hodgkin's lymphoma, Hodgkin's lymphoma, CLL, and multiple myeloma, drawn from Swedish and Danish population-based registries. We used marginal survival methods to test for a relative difference in age at diagnosis between parents and offspring and to account for other risk factors, staggered entries, censored data, and correlations among relatives. Changes in incidence rates of lymphoproliferative tumors were accommodated in the models by using time-varying covariates for different periods of diagnosis. Whereas no anticipation was observed for Hodgkin's lymphoma, CLL, and multiple myeloma, our initial model, which controlled for gender and country, suggested a significant difference (hazard ratio, 0.5; 95% confidence interval, 0.33-0.75) in age at diagnosis between the parents and offspring in the non-Hodgkin's lymphoma sample. However, once we accounted for the significant change in non-Hodgkin's lymphoma incidence over time, the statistical difference between parents and offspring disappeared (hazard ratio, 0.99; 95% confidence interval, 0.56-1.76). Our results emphasize the importance of considering secular trends when evaluating the possibility of anticipation in lymphoproliferative cancers. This is the first study to consider the changes of incidence over time as a source of bias when evaluating anticipation in lymphoproliferative cancers.