To directly compare the diagnostic utility of growth differentiation factor–15 (GDF-15) with our previous fibroblast growth factor–21 (FGF-21) findings in the same adult mitochondrial disease cohort.

Serum GDF-15 levels were measured using a quantitative ELISA. Statistical analyses of GDF-15 data were compared with our published FGF-21 findings.

Median serum GDF-15 concentrations were elevated in patients with mitochondrial disease and differed between all experimental groups, mirroring group results for FGF-21. There was a difference between patients diagnosed by muscle biopsy and genetic diagnosis, suggesting that serum GDF-15 measurement may be more broadly specific for mitochondrial disease than for muscle manifesting mitochondrial disease, in contrast to FGF-21. GDF-15 showed a markedly higher diagnostic odds ratio when compared with FGF-21 (75.3 vs 45.7), was a better predictor of disease based on diagnostic sensitivity (77.8% vs 68.5%), and outperformed FGF-21 on receiver operating characteristic curve analysis (area under the curve 94.1% vs 91.1%). Combining both biomarkers did not improve the area under the curve remarkably over GDF-15 alone. GDF-15 was the best predictor of mitochondrial disease (p < 0.002) following multivariate logistic regression analysis.

GDF-15 outperforms FGF-21 as an indicator of mitochondrial diseases. Our data suggest that GDF-15 is generally indicative of inherited mitochondrial disease regardless of clinical phenotype, whereas FGF-21 seems to be more indicative of mitochondrial disease when muscle manifestations are present.

This study provides Class III evidence that serum GDF-15 accurately distinguishes patients with mitochondrial diseases from those without them.