Neurology -- Correspondence for Wirdefeldt et al., 63 (2) 305-311

Correspondence published:

No evidence for heritability of Parkinson disease in Swedish twins: Michael T. Lin, David K. Simon (4 October 2004)

Reply to Lin et al: Nancy L Pedersen, Karin Wirdefeldt, Margaret Gatz, and Martin Schalling (4 October 2004)

No evidence for heritability of Parkinson disease in Swedish twins 4 October 2004

Michael T. Lin,
Department of Neurology, Weill Medical College of Cornell University
525 East 68th Street, A-501, New York, NY 10021,
David K. Simon
Send Correspondence to journal:
Re: No evidence for heritability of Parkinson disease in Swedish twins

mtl2002{at}med.cornell.edu Michael T. Lin, et al.

In their recent article, Wirdefeldt et al performed a twin study analysis to estimate the contribution of genetics to Parkinson’s disease (PD). [1] They found that monozygotic (MZ) and dizygotic (DZ) twin concordance rates (CR) were low (CRMZ =11%, CRDZ = 8%). The correlations between twins, measured by tetrachoric correlation coefficients (r), were also low (rMZ = 0.39, rDZ = 0.28). None of the differences between MZ and DZ concordances or correlations were statistically significant. Estimates of the fraction of variation due to heredity (H) were only 0.13-0.30, much lower than similar estimates in diseases such as asthma or Alzheimer’s disease. The authors conclude, “our findings indicate that genetic effects are of little importance in PD.” We congratulate the authors on the publication of their thorough study. However, we respectfully disagree that their data exclude an important genetic contribution to PD. While their data do provide evidence that environmental factors influence the risk of PD, this does not necessarily imply that genetic factors do not also play an important role.
We call attention to gene-by-environment interactions, a possibility the authors themselves mention. An incompletely penetrant mutation could reflect gene-by-environment interactions, and we have previously shown that such a mutation could play a major role in determining PD risk, while completely escaping detection by even the most thorough twin study. [2] As an example, suppose there were an autosomal dominant mutation, present in at least one copy in 2.5% of the population, but causing PD in only 10% of subjects carrying at least one copy. If the prevalence of PD were 0.5%, as found by Wirdefeldt et al, the hypothetical mutation would account for PD in 0.25% of the population, representing half of all cases of PD. Moreover, it would increase the risk of disease 39-fold. Such a mutation clearly would have both epidemiologic and biologic importance. However, the low penetrance would result in low probandwise concordance rates (CRMZ = 5.1%, CRDZ = 2.8%), low tetrachoric correlation coefficients (rMZ = 0.36, rDZ = 0.25), small differences between MZ and DZ twins, and small estimates (0.22) for the fraction of variation due to heredity.
These numbers are as small as or even smaller than those found by Wirdefeldt et al, illustrating that their data cannot exclude important genetic contributions to PD risk. Thus, their results are not inconsistent with prior studies indicating a significant genetic contribution to PD. [3-6]
References
1. Wirdefeldt K, Gatz M, Schalling M, Pedersen NL. No evidence for heritability of Parkinson disease in Swedish twins. Neurology 2004;63:305- 311.
2. Simon DK, Lin MT, Pascual-Leone A. "Nature versus nurture" and incompletely penetrant mutations. J Neurol Neurosurg Psychiatry 2002;72:686-689.
3. Marder K, Levy G, Louis ED, et al. Familial aggregation of early- and late-onset Parkinson's disease. Ann Neurol 2003;54:507-513.
4. Payami H, Zareparsi S, James D, Nutt J. Familial aggregation of Parkinson disease: a comparative study of early-onset and late-onset disease. Arch Neurol 2002;59:848-850.
5. Maher NE, Golbe LI, Lazzarini AM, et al. Epidemiologic study of 203 sibling pairs with Parkinson's disease: the GenePD study. Neurology 2002;58:79-84.
6. Sveinbjornsdottir S, Hicks AA, Jonsson T, et al. Familial aggregation of Parkinson's disease in Iceland. N Engl J Med 2000;343:1765- 1770.

Reply to Lin et al 4 October 2004

Nancy L Pedersen,
Karolinska Institutet
Box 281, Stockholm, Sweden,
Karin Wirdefeldt, Margaret Gatz, and Martin Schalling
Send Correspondence to journal:
Re: Reply to Lin et al

Nancy.Pedersen{at}meb.ki.se Nancy L Pedersen, et al.

Lin gives an interesting example of a fairly common (2.5%) autosomal dominant mutation causing Parkinson's disease (PD) only in 10% of individuals. Such a mutation would result in low concordance rates for PD both in monozygotic and dizygotic twins, low tetrachoric correlations as well as a low heritability estimate.
However, we believe that he has over-interpreted the intention of our conclusions. We did not intend to suggest that our findings “exclude important genetic contributions" to PD risk. In contrast, our data indicate that highly penetrant mutations do not play a substantial role in sporadic PD. As Lin et al note, we mention the importance of gene by environment interactions (or incompletely penetrant mutations) as these are difficult to detect using a twin design unless the environmental or genetic factor of interest, or both, is measured. One possible mechanism by which mutations of low penetrance cause disease is increased susceptibility to environmental toxins. In this case, individuals carrying the mutation will only develop the disease once they are exposed to these toxins. We also point out that PET studies suggest a significant genetic influence for nigrostriatial dysfunction. However, these mechanisms appear insufficient to cause clinically detectable PD.
Interpretations of “significant genetic contribution” are value judgements. Lin interprets a penetrance of 10% to represent "significant” contribution. Heritability estimates are relative to total population variation. Any heritability estimate less than 50%, by definition indicates that environmental variance is greater than genetic variance. The model does not provide an estimate of the relative importance of gene by environment interactions. Thus, based on this reasoning and the data we observe, we argue that although susceptibility genes most likely are involved, environmental factors are more important in the etiology of sporadic PD.