As authors of the first paper describing the methodology used by Gialluisi et al in their paper on the high allele frequency for Wilson disease in the Sardinian population, we want to congratulate them with their result.1, 2 The paper clearly shows the strength of this methodology. We were also particularly impressed by their conscientious approach to determine the inbreeding coefficient in this special population.

Still we want to draw attention to some inaccuracies in their paper, which–in our opinion–should be avoided by future users of this method, as they may result in underestimation of the gene frequency – ie, the total pathogenic allele frequency–and birth prevalence of the disorder. First of all, the authors disregard 14 mutations with relative frequencies below 1%. By including these, the gene frequency becomes 0.0195, instead of 0.0191. Second, the authors included only patients with unambiguous genotype and detailed geographical provenance of parents. However, leaving out seemingly heterozygous patients with a second, unidentified mutation will have the same effect on the estimation of the gene frequency as disregarding known mutations. Finally, the authors seem to have calculated the birth frequency of the disorder in Sardinia by simply squaring the gene frequency. This can be justified in random mating populations but not in the Sardinian population where the inbreeding coefficient is higher than zero. For instance in the mountains, where the inbreeding coefficient is 0.00112344, and using the gene frequency estimate of Gialluisi et al2 the birth prevalence of Wilson disease will be 1:2585 instead of the 1:2732 calculated by these authors. Using the gene frequency estimate that includes the 14 rare mutations, the prevalence estimate in the mountains even becomes 1:2499, – 11% higher than the original estimate.

We admit that each of our proposed corrections separately only has a small effect for the final estimate, but these effects are additive, and thus together not always insignificant. Therefore, we hope that by pointing to these inaccuracies future users of the methodology will be forewarned.