Mitochondrial disorders are often fatal multisystem disorders, partially caused by heteroplasmic mitochondrial DNA (mtDNA) point mutations. Prenatal diagnosis is generally not possible for these maternally inherited mutations because of extensive variation in mutation load among embryos and the inability to accurately predict the clinical expression. The aim of this study is to investigate if PGD could be a better alternative, by investigating the existence of a minimal mutation level below which the chance of an embryo being affected is acceptably low, irrespective of the mtDNA mutation. We performed a systematic review of muscle mutation levels, evaluating 159 different heteroplasmic mtDNA point mutations derived from 327 unrelated patients or pedigrees, and reviewed three overrepresented mtDNA mutations (m.3243AG, m.8344AG and m.8993TC/G) separately. Mutation levels were included for familial mtDNA point mutations only, covering all affected (n 195) and unaffected maternal relatives (n 19) from 137 pedigrees. Mean muscle mutation levels were comparable between probands and affected maternal relatives, and between affected individuals with tRNA- versus protein-coding mutations. Using an estimated a priori prevalence of being affected in pedigrees of 0.477, we calculated that a 95 or higher chance of being unaffected was associated with a musc Our data show, for the first time, that carriers of heteroplasmic mtDNA mutations will have a fair chance of having healthy offspring, by applying PGD. Nevertheless, our conclusions are partly based on estimations and, as indicated, do not provide absolute certainty. Carriers of mtDNA should be informed about these constraints.