Response: More on disease anticipation in familial MPN

We thank Rumi et al for commenting on our recently published paper demonstrating strong familial aggregation of MPN.¹  They present their data on ages at diagnosis and telomere lengths in affected parent-offspring pairs in MPN families, which extend their previous report describing apparent “anticipation” in age at diagnosis from the parental to offspring generation.²  The conclusion of “anticipation” in familial diseases is controversial given the well-described truncation bias involved in family studies where the offspring generation is not followed up as long as the parent generation.³  Even in our extensive data obtained by linking Swedish registries covering more than 40 years, the same bias holds due to the start date of the cancer registry (1958) and censoring date of our study (2005). Thus, in our study of familial aggregation of MPN, we concluded that the parental cohort did not have a different age at diagnosis than the offspring as a cohort. As noted by Rumi et al, this is not the same as testing for differences in parent-offspring affected pairs, which we did not present in our paper. In our registry study, we did find 40 families with affected parent-offspring pairs. Not surprisingly the mean age of MPN diagnosis was 68.7 years in parents and 48.7 years in offspring, a highly significant difference. Despite this difference, due to the expected bias due to truncation of data in both parents and offspring, we are reluctant to conclude that there is anticipation.

Rumi et al have also presented data showing that the MPN offspring have shorter telomeres (relative to their age group) than do the affected parents. One can hypothesize that shorter telomeres could be a reflection of greater genomic instability leading to cancer at an earlier age and could be a result of some epigenetic alteration from parent to offspring. However, there are certain methodologic limitations in these measures due to factors such as substantial interindividual variation (even within age groups) and inherited variation of telomere length.^4-6  It would be important to compare telomere lengths not just between affected parents and offspring but between the affecteds and unaffecteds within the same families.⁷  Future studies of familial MPN should help clarify these issues and unravel the relevant pathways leading to both familial and sporadic MPN.