Mitochondrial diabetes mellitus: real world insights from the GENOMIT registry-a multinational, longitudinal cohort study

EBioMedicine. 2026 Sep 1;131:106458. doi: 10.1016/j.ebiom.2026.106458. Online ahead of print.

ABSTRACT

BACKGROUND: Diabetes mellitus is a common but incompletely characterised manifestation of mitochondrial diseases (MD). Data on risk factors, clinical course, and treatment recommendations are lacking.

METHODS: In this multinational cohort study, we analysed longitudinal data of patients with a genetically confirmed MD from the GENOMIT registry included at German, Austrian, and Italian sites between 07/2009-01/2025. Our objectives were to (1) expand the genetic spectrum of mitochondrial diabetes mellitus (mDM), (2) identify risk factors, (3) delineate the clinical course, and (4) characterise real-world use of antidiabetic therapies.

FINDINGS: Of 2399 patients, 1225 (51%) were female, and 281 (12%; 172 female) had mDM. Diabetes occurred across 31 genotypes and exhibited marked genotype dependence, with the highest prevalence in m.3243A>G carriers (177/360 [49%]). Only the m.3243A>G variant was associated with a significantly increased risk of mDM (HR = 10.3; 95% CI 5.2-20.4, p < 0.0001), whereas single mtDNA deletions, multiple mtDNA deletions, and primary LHON variants, as well as sex, BMI, ethnicity, smoking, hypertension and dyslipidaemia did not show a significant association. Median diabetes onset in patients with the m.3243A>G variant was at 47.7 years (SD 45.2-52.0). Among patients with mDM, 140/281 (50%) used insulin, and 111/281 (40%) received non-insulin antidiabetic drugs, most commonly metformin, which was discontinued in 8/50 users. Literature review revealed neurological events temporally linked to metformin application in m.3243A>G carriers, though long-term use without adverse events was likewise reported.

INTERPRETATION: mDM is frequent in patients with MD, and the individual risk is strongly genotype dependent. While caution is warranted, our data do not justify universal avoidance of metformin; prospective, genotype-informed studies are needed to guide management.

FUNDING: German Ministry of Research, Technology and Space; Italian Ministry of Health; European Union.

PMID:42679752 | DOI:10.1016/j.ebiom.2026.106458