Anticipation in familial Parkinson's disease: a reanalysis of 13 United Kingdom kindreds.

Several authors reported anticipation for age at onset in familial Parkinson's disease (PD) and postulated the involvement of an unstable trinucleotide repeat gene. We reanalyzed 13 previously reported multiple generation kindreds with familial PD to investigate anticipation and related biases. Although initial review of 33 informative parent-child and aunt/uncle-niece/nephew pairs revealed a mean anticipation interval of 9.3 years (p = 0.0001), exclusion of probands, in an effort to limit referral bias, reduced this to 4.1 years (p = 0.09). This reduced figure is still an overestimate of the true anticipation interval because of residual referral bias and because of incomplete information on the most recent generation. Our analyses did not demonstrate the effects of recall bias or of reduced fertility. We conclude that the observed anticipation in familial PD may partly or completely result from referral and other biases. Caution should be exercised in assuming an unstable trinucleotide repeat gene as the cause of familial PD.