With interest we read the article by Maresca et al. about a study of fibroblast growth factor 21 (FGF21), growth differentiation factor 15 (GDF-15), and the cell free circulating mitochondrial DNA (ccf-mtDNA) as potential biomarkers for mitochondrial disorders (MIDs) [1]. The study population comprised 123 MID patients. It was concluded that FGF21, GDF-15, are associated with mitochondrial translation defects due to tRNA mutations [1]. Additionally, ccf-mtDNA was strongly associated with mitochondrial encephalopathy, lactic acidosis, and stroke-like episode (MELAS) syndrome, why it was proposed to use it for monitoring the disease course and the treatment effect in the acute phase [1]. We have the following comments and concerns.

For calculating sensitivity and specificity of a biomarker it is crucial that a disease control group is investigated. However, only 15 healthy controls and 123 MID patients were investigated. We should know which cohort of patients was used as a disease control group. 

The senitivity of the three molecules investigated was low. The sensitivity of ccf-mtDNA was calculated as only 25%, that of FGF21 as 42%, and that of GDF-15 as 89%. This means that the number of false positive results was high. 

Which is the reason why the sensitivity of the molecules investigated was low? The low sensitivity of ccf-mtDNA could be explained by the fact that ccf-mtDNA can be also elevated in bipolar depression [2] or in patients with coronary heart disease and diabetes [3], Additionally, decreased ccf-mtDNA in the cerebro-spinal fluid (CSF) is related to progressive cell dysfunction in neurodegenerative disorders, such as Alzheimer’s disease, and Parkinson’s desease [4]. Ccf-mtDNA can be elevated in the CSF of patients with multiple sclerosis [4]. There are also indications that ccf-mtDNA contributed to chronic inflammation of diabetic patients [5]. 

The low sensitivity of FGF21 could be explained by non-specific elevation of FGF-21 in a number of other conditions, such as preterm newborns [6], in patients with TK2-related myopathy [7], in patients with non-alcoholic fatty liver disease [8], atherosclerosis [9], and many other conditions. 

Elevation of GDF-15 has been reported not only in MIDs but also in patients with multisystem atrophy [10], patients with metastasising cancer [11], and patients with reduced muscle mass [12]. Since GDF-15 correlates with muscle mass, we should know how many of the included patient had muscle wasting and elevated GDF-15. 

Since the phenotype of mtDNA variants may not only depend on heteroplasmy rates but also on the mtDNA copy number, we should know if mtDNA copy correlated significantly with any of the three molecules tested as a potential biomarker.

Overall, this interesting study has a number of shortcomings which do not allow currently to draw conclusions as done and request provision of additionally data. Particularly, a disease control group needs to the implemented in the study. If a disease control group is investigated which is known to develop high levels of these three molecules, the sensitivity of the three molecules will most probably further decline. Thus, we doubt that GDF-15, ccf-mtDNA, and FGF21 can serve as biomarkers for MIDs.