In type I familial amyloid polyneuropathy (FAP) caused by a variant Met30-transthyretin (TTR), genetic anticipation has been reported. To determine whether anticipation of the disease is a true biological phenomenon or the result of ascertainment bias, we compared age at onset of the affected child with that of the affected parent in 68 parent-child pairs (including data on assumed age at onset and on asymptomatic obligate heterozygotes and parents at obligate 50% risk) in 15 families. Excluding the parent-child pairs involving the proband and "bilineal pairs", onset occurred earlier in the child than in the transmitting parent in 60 out of 68 "unilineal pairs". After correction for ascertainment bias resulting from incomplete penetrance and reduced biological fitness in early onset patients, the number of anticipation pairs (60 pairs) was still significantly larger than that of non-anticipation pairs (29.7 pairs) (p < 0.05). When the children were sons, the difference in age at onset was significantly greater in the mother-son pairs than in the father-son pairs (p = 0.023). Although not all ascertainment biases could be eliminated, these data show strong evidence that anticipation in the transmission of Met30-TTR FAP is a true biological phenomenon.