The accidental amplification of nuclear mitochondrial pseudogenes (NUMTs) can pose a serious problem for mitochondrial disease studies. This report shows that the mutation spectrum left by spurious amplification of a NUMT can be detected because it usually differs considerably from the authentic natural spectrum. This study examined the problem introduced by an ND5 gene NUMT that was recorded in a proband with hearing loss and reviews other disease studies erroneously reporting NUMT variation as genuine mutations in their patients. NUMTs can emerge in population genetic studies, as exemplified here by cases in this study and from published sources. Appropriate database searches and a phylogenetic approach can prevent hasty claims for novelty of mitochondrial DNA (mtDNA) variants inadvertently derived from NUMTs and help to direct investigators to the real source.