In families with one affected child, mosaicism of pa­rental lymphocytes or fibroblasts is found in under 5% of cases [26]. It is more common in families with DS recur­rence, but even then, the incidence is relatively low. For example, one study found mosaicism in only five out of 22 families, even using DNA analysis which, unlike cyto­genetic methods, may detect low-level mosaicism [19]. There are case reports of gonadal mosaicism with a normal chromosome complement in peripheral blood or fibro­blasts, but the proportion of families with such errors is not known. However, even a high percentage of trisomic oo­cytes cannot explain all recurrences: when all are trisomic, half the gametes would be normal. And a serious criticism of all published mosaicism studies is that they did not include controls from unaffected families. Studies on pre-implantation genetic diagnosis show that aneuploidy in oocytes and embryos is not a rare event and it increases markedly with maternal age, being the consequence of both trisomic germ line and further disruption in meiotic division [27].

Although there is evidence for mtDNA involvement in DS pathogenesis, it is widely believed that the oxidative stress in DS is due to a direct gene dosage effect resulting from a 50% increase in the activity of SOD-1. However, examination of the available evidence reveals many con­tradictions [28].

The association between trisomy 21 and SOD-1 activ­ity is not straight-forward: some DS cases with partial trisomy 21, a translocation or mosaic replication, have normal SOD-1; some cases of partial trisomy 21 without DS have increased SOD-1 [8,29]; and in some DS indi­viduals with partial trisomy 21 the SOD-1 locus is not triplicated [30]. In DS cases, SOD-1 levels vary according to cell or tissue type: in blood cells and fibroblasts activity is generally increased, albeit to a variable extent, while in some individuals who have normal activity and plasma SOD‑1, activity is markedly decreased [29,31-34]. Fur­thermore, it is far from clear that increased SOD‑1 in DS is a cause of oxidative stress. In one series of DS cases it was found that while SOD‑1 levels fell with age, lipid peroxidation increased [9,12]. An alternative and compel­ling interpretation of the data on elevated SOD‑1 activity is that it has a protective rather than a harmful effect, since the increase of this scavenger is of benefit for the homeostasis between generated reactive oxygen metabo­lites and their propagation. In this interpretation SOD‑1 activity in DS is increased as a response to oxidative stress, and unlike unaffected individuals, activity can readily be increased because of the additional copy of the gene. In conclusion, there is more evidence in favor of a role for mtDNA mutations in both DS etiology and patho­genesis than for any of the alternative explanations.