The mitochondrial disease group consists of different disorders with unprecedented variability of clinical manifestations and tissue-specific symptoms. Their tissue-specific stress responses vary depending on the patients' age and type of dysfunction. These responses include secretion of metabolically active signal molecules to systemic circulation. Such signals-metabolites or metabokines-can be also utilized as biomarkers. During the past 10 years, metabolite and metabokine biomarkers have been described for mitochondrial disease diagnosis and follow-up, to complement the conventional blood biomarkers lactate, pyruvate and alanine. These new tools include metabokines FGF21 and GDF15; cofactors (NAD-forms); sets of metabolites (multibiomarkers) and the full metabolome. FGF21 and GDF15 are messengers of mitochondrial integrated stress response that together outperform the conventional biomarkers in specificity and sensitivity for muscle-manifesting mitochondrial diseases. Metabolite or metabolomic imbalance (e.g., NAD+ deficiency) is a secondary consequence to the primary cause in some diseases, but relevant as a biomarker and a potential indicator of therapy targets. For therapy trials, the optimal biomarker set needs to be tailored to match the disease of interest. The new biomarkers have increased the value of blood samples in mitochondrial disease diagnosis and follow-up, enabling prioritization of patients to different diagnostic paths and having crucial roles in follow-up of therapy effect.