eISSN: 2083-8441
ISSN: 2081-237X
Pediatric Endocrinology Diabetes and Metabolism
Current issue Archive Manuscripts accepted About the journal Supplements Editorial board Reviewers Abstracting and indexing Subscription Contact Instructions for authors Ethical standards and procedures
SCImago Journal & Country Rank
vol. 28
Original paper

The fibroblast growth factor 21 concentration in children with mitochondrial disease does not depend on the disease stage, but rather on the disease genotype

Dorota Wesół-Kucharska
Dariusz Rokicki
Milena Greczan
Magdalena Kaczor
Edyta Czekuć-Kryśkiewicz
Dorota Piekutowska-Abramczuk
Paulina Halat-Wolska
Elżbieta Ciara
Maciej Jaworski
Aleksandra Jezela-Stanek

Department of Pediatrics, Nutrition and Metabolic Diseases, The Children’s Memorial Health Institute, Warsaw, Poland
Laboratory of Radioimmunology and Experimental Medicine, Department of Biochemistry, Radioimmunology and Experimental Medicine; The Children’s Memorial Health Institute, Warsaw, Poland
Department of Medical Genetics, The Children’s Memorial Health Institute, Warsaw, Poland
Department of Biochemistry, Radioimmunology and Experimental Medicine, The Children’s Memorial Health Institute, Warsaw, Poland
Department of Genetics and Clinical Immunology, National Institute of Tuberculosis and Lung Diseases, Warsaw, Poland
Pediatr Endocrinol Diabetes Metab 2022; 28 (2): 141–151
Online publish date: 2022/05/12
Article file
Get citation
JabRef, Mendeley
Papers, Reference Manager, RefWorks, Zotero


Mitochondrial diseases (MD) are a heterogeneous group of disorders resulting from a damaged function of the mitochondrial respiratory chain and impaired ATP production in the process of oxidative phosphorylation. Mitochondrial diseases are rare diseases but their total incidence is 1.6 : 5000 of live births, which makes them the most common diseases among inherited metabolic disorders [1, 2]. The clinical picture of MD is highly varied. The disease may be manifested at any age and virtually each organ may manifest little specific symptoms. The most common symptoms are related to high-energy organs: central nervous system (encephalopathy, seizures, stroke-like lesions, psychomotor developmental delay), skeletal muscles (myopathy, exercise intolerance), sensory organs (sensory-neural hearing loss, optic nerve atrophy, ptosis, external ophthalmoplegia), but they may also involve the heart (hypertrophic cardiomyopathy, arrhythmia), with possible endocrine disorders or gastrointestinal symptoms [3, 4]. These diseases may be related to an abnormal variant in mitochondrial DNA (mtDNA) or nuclear DNA (nDNA). Up to now, over 400 genes responsible for MD have been recognised [1, 3, 5].
Due to a very high heterogeneity of clinical phenotypes occurring even in subjects with damage to the same causative gene, clinical evaluation of MD patients should be based on various clinical scales (i.a. the International Paediatric Mitochondrial Disease Scale, IPMDS) [6], as well as specific biomarkers (i.a. FGF21). The scale includes a subjective assessment (problems reported by guardians), physical examination with elements of neurological examination, and a functional assessment. It is used to assess MD severity and stage in children (0–18 years), and it may also be used to assess MD progression [6–8]. FGF21 has been shown to be a sensitive and specific biomarker of MDs, especially those with skeletal muscle involvement [9–11]. The increased FGF21 concentration in MD is correlated with impaired oxidative phosphorylation and reduced ATP production, which means metabolic stress to the cells, to which they respond with FGF21 production [12, 13].
FGF21 may also be used as a marker of the MD stage; higher and gradually increasing concentrations have been observed in subjects in the final stage of the disease and with higher damage to skeletal muscles (FGF21 concentration was correlated with a higher percentage of COX-negative fibres in the muscle biopsy) [9, 14, 15].
The study objective was to verify if the FGF21 concentration in paediatric patients with mitochondrial disease was correlated with the disease severity and stage.

Material and methods

Paediatric patients with a probable, or molecularly confirmed mitochondrial disease were enrolled in the study. MD probability was assessed with the use of the Nijmegen scale (subjects with a score higher than 4 were included in the study group) [16, 17]. Molecular tests were performed depending on clinical preconditions: MLPA analysis and/or Sanger sequencing of selected mtDNA fragments or an authorial NGS (Next Generation Sequencing) panel for identification of known mitochondrial diseases depending on the presence of nDNA/mtDNA changes or for identification of common point mutations and long-range mtDNA rearrangements.
Each MD patient was assessed by researchers using IPMDS according to the guidelines shared by the scale authors [7]. The result is given as the percentage of the score that may be achieved by a given patient. The patients were assessed in 3 domains, the scores were added and the total value was calculated (IPMDS total). The patients with the IPMDS total < 10% were qualified as mild MD phenotype (group A), IPMDS total = 10–19,9% were qualified as moderate severity MD (group B) and with IPMDS total ≥ 20% – as severe MD (group C).
The control group comprised of children without obesity (BMI < 95th pc for age and sex), without signs of liver dysfunction, nephropathy or circulatory insufficiency (since these are known conditions with potentially elevated FGF21). Among the control group: 7 were healthy children (mainly siblings of study patients), 5 – patients with delayed psychomotor development, without confirmed inborn error of metabolism (IEM), 8 – had drug-resistant epilepsy (5 with molecularly confirmed IEM: SLC2A1, MOCS2, GLUL, ADSL) and 1 – optic nerve atrophy (without IEM).
FGF21 was analysed in plasma samples using the BioVendor Human FGF-21 ELISA (BioVendor – Laboratorní medicína a.s., Brno, Czech Republic). Together with FGF21, plasma concentrations of conventional biomarkers lactate (LA), pyruvate (PA), alanine and creatine kinase (CK) were measured. Statistical differences were determined using Kruskal–Wallis and ANOVA non-parametric tests for data with no normal distribution (FGF21 concentration) and ANOVA parametric test – for data whose distribution was no different from the normal distribution (IPMDS). A p-value of < 0.05 was considered statistically significant. Correlations were assessed with Spearman’s correlation test.
The study was approved by the Committee for Medical Research Ethics of Children’s Memorial Health Institute in Warsaw (No: 40/KBE/2019), and written consent was obtained from all patients/ parents of patients.


Thirty-two patients with MD and 21 children without MD were assessed. The median age in the study group and in the control group was 33 (2–213) and 42 (8–202) months, respectively. The clinical characteristics of the study group are presented in Table I. In 6 (18.7%) patients the molecular cause of MD was not identified, in 3 patients the diagnosis was made on the basis of selective screening – testing relatives (in those patients the Nijmegen score at diagnosis was < 4).
Regarding severity of mitochondrial disease, assessed using IPMDS, patients with MD were divided into 3 groups. In group A (n = 10), with the least severe symptoms, the patients were older than in the other groups (group B – with moderate severity, n = 9), or group C (severe MD phenotype, n = 13), whereas in group C the disease had the earlies onset – the data are presented in Figure 1. In patients from groups A and B epileptic seizures were less frequent (epilepsy was diagnose in 3/10 and 1/9 patients, respectively) than in group C, where seizures occurred in 6/13 patients, although the correlation was not statistically significant. Developmental delay was significantly more frequent in patients with a more advanced disease. In groups A, B and C it was observed in 4/10, 7/9 and 13/13 subjects, respectively.
The concentration of FGF21, lactic acid, pyruvic acid, alanine, and creatine kinase in particular groups are shown in Table II. The FGF21 concentration in the study MD group was significantly higher (p < 0.01) than in the control group (576.5 vs 30.9 pg/ml), and the cut-off point for MD FGF21 (> 95th pc) was 266 pg/ml. Similarly, the concentration of LA (34.2 vs 13.9 mg/dl), PA (1.4 vs 0.9 mg/dl) and the lactic acid to pyruvic acid ratio (LA/PA) (20.1 vs 13.3) were higher in the group with MD. No significant correlation was observed between MD stage measured by IPMDS and the FGF21 concentration (Rs = 0.0032, p = 0.9857) (Fig. 2). LA and PA concentrations were the highest in group C with the severe MD phenotype, and the lowest in patients with mild symptoms (group A). The difference, however, was not statistically significant. No such correlations were observed for other MD biomarkers, such as LA/PA ratio, serum alanine or CK.
Genetic diagnosis could not be made in 6 patients with MD. In 10 patients an abnormal variant was found in mtDNA (mtMD group) in genes: MT-TL1 (n = 4), MT-ATP6 (n = 2), MT-TK1 (n = 1), MT-ND3 (n = 1), MT-ND5 (n = 1), MT-ND6 (n = 1); in 7 patients an abnormal variant was found in PDHA1 gene (group PMD), and in 9 patients in other nDNA genes (group nMD): SURF1 (n = 2) and in the following genes damaged in one patient each: SCL19A3, SLC25A12, COQ8A, EARS2, FLBX4, NDUFS1, NUBPL. Biochemical data for particular groups of patients are shown in Table II. Patients in nMD and PMD groups were younger (the median age was 16 and 40 months, respectively) and showed an earlier onset of the disease (median age: 6 months in both groups) than patients with mtMD damage (median current age: 100 months, median first symptoms: 40 months). Patients from group nMD had a more advanced disease (IPMDS total = 21.4 %) as compared to PMD (IPMDS total = 8.6%) and mtMD patients (IPMDS total = 12.5%), but the difference was not statistically significant (p = 0.1190) (Figure 3). The FGF21 concentration was the highest in group nMD (median FGF 21 = 1022.0 pg/ml) and the lowest in PMD patients (median FGF 21 = 28.2 pg/ml), and the difference was significant (p < 0.05) (Fig. 4). In patients with disorders of mitochondrial translation, including abnormal variants in mitochondrial tRNA (MT-TL1, MT-TK, EARS2, FBXL4) FGF21 concentration was higher (median FGF21 = 2317 pg/ml, range: 751–8873) than in patients with abnormal genetic variants responsible for the subunit structure of the respiratory chain (MT-ND3, MT-ND5, MT-ND6, MT-ATP6, SURF1, NDUFS1, NUBPL) (median FGF21 = 413 pg/ml, range: 106–1022).
The LA concentration was also the highest in the group of nMD patients (median: 42.5 mg/dl). The PA concentration was the highest in group PMD (median PA = 1.7 mg/dl), and this group had the lowest LA/PA (median LA/PA = 14.5) – which results from biochemical specification of the disease. There was a slight correlation between FGF21 and the concentration of LA, LA/PA and serum alanine in MD patients (Figs. 5–7).


FGF21 is a protein produced primarily by the liver in response to the state of prolonged fasting. Its expression is also demonstrated in other tissues, e.g. in the adipose tissue, skeletal muscles, pancreas, cardiac endothelial cells, or in the hypothalamus [12]. An increased FGF21 concentration was observed in various conditions related to metabolic stress (e.g. after prolonged exercise), but also in quite common diseases – e.g. non-alcoholic fatty liver disease, obesity, undernutrition, insulin resistance, coronary heart disease, or nephropathy [12, 18, 19]. Its concentration is not correlated with age or sex [9, 20, 21]. Also, there is no correlation between the FGF21 concentration and liver dysfunction [9, 20, 21]. FGF21 sensitivity and specificity to mitochondrial myopathies has appeared to be higher than that of other known biochemical markers, such as: serum lactic acid concentration, serum lactic acid to pyruvate ratio, assessment of serum alanine concentration [9, 14, 20, 22, 23]. FGF21 has appeared to be a more sensitive index than muscle biopsy; a significant increase in the FGF21 concentration has been observed in patients with mitochondrial disease confirmed by a molecular examination, in whom muscle biopsy did not show changes [24].
In the presented study, FGF21 has appeared to be a sensitive and specific MD marker. The usefulness of this protein, as new biomarkers for MD diagnosis has been shown by numerous study [9, 10, 14, 20, 25–27] and confirmed by meta-analysis [22]. FGF21 sensitivity in the meta-analysis referred to was: from 0.51 to 0.93 (I2 91.26%, 95% CI: 85.23–97.29) and from 0.77 to 0.97 for specificity (I2 66.31%, 95% CI: 34.06–98.57) across the five studies, with a total of 718 participants (339 patients and 379 healthy controls) [22]. A cut-off point (> 95th pc) for the MD concentration (vs healthy control) given in the literature is from 180 to 400 pg/ml, and the values were several/several dozen times higher (2.2–13 times) than in the control group [9–11, 14, 20, 21]. Similar data have been achieved in the present study.
FGF21 have turned out to be more sensitive and specific than other known biomarkers MD (LA, PA, LA/PA, CK, or serum alanine) [21, 26, 28]. No correlation has been found between FGF21 and the other MD markers [21, 26], the opposite data result have been achieved in the conducted study, but it was a slight correlation. FGF21 have been suggested to be used for monitoring of MD treatment [21, 26, 29].
There are no consistent data, however, showing a correlation between MD stage and the concentration of FGF21. Yatsuga et al. [27] showed a positive correlation of FGF21 with the disease stage in the group of 48 patients with MD, (median age: 33.6 ±18.7 years) with a different phenotype and genotype. Similar data were achieved by Koene et al. [30] in the assessment of adult patients with a common pathogenic variant in mtDNA m.3243A>G. Davis et al. [10] did not achieve such a correlation in a group of adults with MD (different disease phenotype and genotype) – no higher FGF21 concentrations were observed in subjects with a more advanced MD. The present study has also showed no such correlation in paediatric patients with a different disease phenotype and genotype – the FGF21 concentration was not higher in children with more severe symptoms of MD.
There are no consistent data that would show an increase in FGF21 with MD progression. In their study on mice, Tyynismaa et al. [15] observed that FGF21 was increasing together with the progression of muscle disease in animals with mtDNA deletion and respiratory chain deficit. Soumalainem et al. [9] observed increasing values of FGF21 in several MD patients who at the same time showed an increase in the severity of clinical symptoms. Similarly, Maresca et al. [28] recorded significantly higher FGF21 concentration in 4 patients with MD during a stroke-like episode (SLE), as compared to the concentration from before the disease exacerbation. On the other hand, Koene [30] did not observe increasing FGF21 concentrations during a 24-month follow-up of patients with MD caused by m.3243A>G change.
Numerous studies have demonstrated high FGF21 concentrations in mitochondrial depletion syndromes [14, 20, 21] and in disorders of mitochondrial translation, including abnormal variants in mitochondrial tRNA [20, 21, 25, 28]. In 26 adult patients with MD, higher FGF21 concentrations were observed if the disease was caused by an abnormal variant in mtDNA (in MD caused by a single mutation in mtDNA (FGF21 = 730 pg/ml) and mtDNA deletion (FGF21 = 1172 pg/ml), respectively, in comparison with subjects whose MD resulted from an abnormal variant in nDNA (FGF21 = 401 pg/ml), but the difference was not statistically significant [25]. Similar data were achieved by Maresca et al. [28] in a group of MD adults n = 123: significantly higher FGF21 values were observed in patients with abnormal mtDNA variant, especially in patients with MELAS and MERRF phenotypes. In paediatric patients [21], however, an inverse correlation was observed – lower FGF21 concentrations occurred in children with MD caused by a change in mtDNA (n = 24; median: 147 (30–4300) pg/ml), as compared to the children whose disease resulted from a change in nDNA (n = 24; median: 362 (30–6200 pg/ml). Similar observations follow from the present paper. There is too little research to explain this finding. It is known that in children MD more often results from an abnormal variant in nDNA, while in adults it results from mtDNA [3, 31], where the most frequently described change is m.3243A>G, in which high levels of FGF21 are observed [28, 31]. Early onset of MD is also associated with a worse prognosis and a more severe course, which may be the reason why FGF21 levels are higher in paediatrics patients with a pathogenic nDNA variant than in children with mtDNA alteration. The lowest levels of FGF21, often below the cut-off point for MD, were observed in patients with an abnormal variant responsible for the structure of the respiratory chain subunits [14, 20]. An interesting fact is that children with an abnormal variant in the PDHA1 gene show often normal values of the above parameters (in 4/ 5 patients they were < 95th pc) [21]. Our observations are similar: in the group of 7 children with a change in the PDHA1 gene, the FGF21 concentrations were lower than in other MD patients, and in 5/7 patients FGF21 was < 95th pc.


The FGF21 concentration is significantly higher in paediatric patients with MD than in children without MD. There is no correlation between the stage of MD and FGF21 concentration. Higher FGF21 values are seen in patients whose MD results from an abnormal nDNA variant rather than mtDNA damage, whereas lower values, often below the MD cut-off point, are seen in the deficit of pyruvic acid dehydrogenase.


1. Stenton SL, Prokisch H. Genetics of mitochondrial diseases: Identifying mutations to help diagnosis. EBioMedicine 2020; 56: 102784. doi: 10.1016/j.ebiom.2020.102784.
2. Tan J, Wagner M, Stenton SL, et al. Lifetime risk of autosomal recessive mitochondrial disorders calculated from genetic databases. EBioMedicine 2020; 54: 102730. doi: 10.1016/j.ebiom.2020.102730.
3. Coelho MP, Martins E, Vilarinho L. Diagnosis, management, and follow-up of mitochondrial disorders in childhood: a personalized medicine in the new era of genome sequence. Eur J Pediatr 2019; 178: 21–32. doi: 10.1007/s00431-018-3292-x.
4. Davison JE, Rahman S. Recognition, investigation and management of mitochondrial disease. Arch Dis Child 2017; 102: 1082–1090. doi: 10.1136/archdischild-2016-311370.
5. Parikh S, Goldstein A, Karaa A, et al. Patient care standards for primary mitochondrial disease: a consensus statement from the Mitochondrial Medicine Society. Genet Med 2017; 19: 10.1038/gim.2017.107. doi: 10.1038/gim.2017.107.
6. Goldstein A, Rahman S. Seeking impact: Global perspectives on outcome measure selection for translational and clinical research for primary mitochondrial disorders. J Inherit Metab Dis. 2020; 44: 343–357. doi: 10.1002/jimd.12320.
7. Koene S, Hendriks JCM, Dirks I, et al. International Paediatric Mitochondrial Disease Scale. J Inherit Metab Dis 2016; 39: 705–712. doi: 10.1007/s10545-016-9948-7.
8. Mancuso M, McFarland R, Klopstock T, Hirano M and consortium on Trial Readiness in Mitochondrial Myopathies. International Workshop: Outcome measures and clinical trial readiness in primary mitochondrial myopathies in children and adults. Consensus recommendations.16-18 November 2016, Rome, Italy. Neuromuscul Disord 2017; 27: 1126–1137. doi: 10.1016/j.nmd.2017.08.006.
9. Suomalainen A, Elo JM, Pietiläinen KH, et al. FGF-21 as a biomarker for muscle-manifesting mitochondrial respiratory chain deficiencies: a diagnostic study. Lancet Neurol 2011; 10: 806–818. doi: 10.1016/S1474-4422(11)70155-7.
10. Davis RL, Liang C, Sue CM. A comparison of current serum biomarkers as diagnostic indicators of mitochondrial diseases. Neurology 2016; 86: 2010–2015. doi: 10.1212/WNL.0000000000002705.
11. Suomalainen A. Fibroblast growth factor 21: a novel biomarker for human muscle-manifesting mitochondrial disorders. Expert Opin Med Diagn 2013; 7: 313–317. doi: 10.1517/17530059.2013.812070.
12. Gómez-Sámano MÁ, Grajales-Gómez M, Zuarth-Vázquez JM, et al. Fibroblast growth factor 21 and its novel association with oxidative stress. Redox Biol 2016; 11: 335–341. doi: 10.1016/j.redox.2016.12.024.
13. Forsström S, Jackson CB, Carroll CJ, et al. Fibroblast Growth Factor 21 Drives Dynamics of Local and Systemic Stress Responses in Mitochondrial Myopathy with mtDNA Deletions. Cell Metab 2019; 30: 1040–1054.e7. doi: 10.1016/j.cmet.2019.08.019.
14. Formichi P, Cardone N, Taglia I, et al. Fibroblast growth factor 21 and grow differentiation factor 15 are sensitive biomarkers of mitochondrial diseases due to mitochondrial transfer-RNA mutations and mitochondrial DNA deletions. Neurol Sci 2020; 41: 3653–3662. doi: 10.1007/s10072-020-04422-5.
15. Tyynismaa H, Carroll CJ, Raimundo N, et al. Mitochondrial myopathy induces a s tarvation-like response. Human Molecular Genetics 2010; 19: 3948–3958. doi: 10.1093/hmg/ddq310.
16. Morava E, van den Heuvel L, de Vries MC, et al. Mitochondrial disease criteria. Diagnostic applications in children. Neurology 2006; 67: 1823–1826. doi: 10.1212/01.wnl.0000244435.27645.54.
17. Pronicka E, Piekutowska-Abramczuk D, Pronicki M. Choroby mitochondrialne u dzieci - podłoże biochemiczne i molekularne, ze szczególnym uwzględnieniem zespołu Leigha. Postępy Biochemii 2008; 54: 161–168.
18. Zhang X, Yeung DC, Karpisek M, et al. Serum FGF21 levels are increased in obesity and are independently associated with the metabolic syndrome in humans. Diabetes 2008; 57: 1246–1253. doi: 10.2337/db07-1476.
19. Tyynismaa H, Raivio T, Hakkarainen A, et al., Liver fat but not other adiposity measures influence circulating FGF21 levels in healthy young adult twins. J Clin Endocrinol Metab 2011; 96: E351–E355. doi: 10.1210/jc.2010-1326
20. Lehtonen JM, Forsström S, Bottani E, et al. FGF21 is a biomarker for mitochondrial translation and mtDNA maintenance disorders. Neurology 2016; 87: 2290–2299. doi: 10.1212/WNL.0000000000003374.
21. Tsygankova PG, Itkis YS, Krylova TD, et al. Plasma FGF-21 and GDF-15 are elevated in different inherited metabolic diseases and are not diagnostic for mitochondrial disorders. J Inherit Metab Dis 2019; 42: 918–933. doi: 10.1002/jimd.12142.
22. Lin Y, Ji K, Ma X, et al. Accuracy of FGF-21 and GDF-15 for the diagnosis of mitochondrial disorders: A meta-analysis. Ann Clin Transl Neurol 2020; 7: 1204–1213. doi: 10.1002/acn3.51104.
23. Davis RL, Liang C, Edema-Hildebrand F, et al. Fibroblast growth factor 21 is a sensitive biomarker of mitochondrial disease. Neurology 2013; 81: 1819–1826. doi: 10.1212/01.wnl.0000436068.43384.ef.
24. Lehtonen JM, Auranen M, Darin N, et al. Diagnostic value of serum biomarkers FGF21 and GDF15 compared to muscle sample in mitochondrial disease. J Inherit Metab Dis 2021; 44: 469–480. doi: 10.1002/jimd.12307
25. Varhaug KN, Hikmat O, Nakkestad HL, et al. Serum biomarkers in primary mitochondrial disorders. Brain Commun 2021; 3: fcaa222. doi: 10.1093/braincomms/fcaa222.
26. Montero R, Yubero D, Villarroya J, et al. GDF-15 is elevated in children with mitochondrial diseases and is induced by mitochondrial dysfunction. PLoS One 2016; 11: e0148709. doi: 10.1371/journal.pone.0148709.
27. Yatsuga S, Fujita Y, Ishii A, et al. Growth differentiation factor 15 as a useful biomarker for mitochondrial disorders. Ann Neurol 2015; 78: 814–823. doi: 10.1002/ana.24506.
28. Maresca A, Del Dotto V, Romagnoli M, et al. ER-MITO Study Group. Expanding and validating the biomarkers for mitochondrial diseases, J Mol Med (Berl) 2020; 98: 1467–1478. doi: 10.1007/s00109-020-01967-y
29. Koga Y, Povalko N, Inoue E, et al. Biomarkers and clinical rating scales for sodium pyruvate therapy in patients with mitochondrial disease. Mitochondrion 2019; 48: 11–15. doi: 10.1016/j.mito.2019.02.001.
30. Koene S, de Laat P, van Tienoven DH, et al. Serum FGF21 levels in adult m.3243A>G carriers: clinical implications. Neurology 2014; 83: 125–133. doi: 10.1212/WNL.0000000000000578.
31. Barca E, Long Y, Cooley V, et al. Mitochondrial diseases in North America: An analysis of the NAMDC Registry. Neurol Genet 2020; 6: e402. doi: 10.1212/NXG.0000000000000402.

Quick links
© 2022 Termedia Sp. z o.o. All rights reserved.
Developed by Bentus.