Clinical spectrum and genetic causes of mitochondrial hepatopathy phenotype in children

**Background:** Mitochondrial hepatopathies (MH) result from primary dysfunction of hepatocyte mitochondria leading to cell injury, steatosis, or liver failure. They are increasingly recognized causes of pediatric liver disease, particularly acute liver failure in infants under 2 years. The genetic basis involves both mitochondrial DNA (mtDNA) and nuclear DNA (nDNA) genes, with over 300 genes linked to mitochondrial disorders. This study aimed to identify genetic causes, characterize clinical phenotypes, and determine outcomes in the largest prospective cohort of children with MH reported to date.

**Methods:** Data were analyzed from 86 unique participants: 64 consecutive patients enrolled in the Longitudinal Study of Mitochondrial Hepatopathy (MITOHEP) and 22 from the Pediatric Acute Liver Failure Study Group (PALFSG) classified as MH. Participants were divided into three groups: ALF (n=37, defined by PALFSG criteria), chronic liver disease (n=40, defined by liver dysfunction >6 months or abnormal histology), and post-liver transplant (n=9). Genetic testing included a targeted gene panel of 26 nuclear genes and whole exome sequencing. Respiratory chain enzyme testing was performed on liver, muscle, or skin fibroblasts. Statistical comparisons used chi-square or Wilcoxon tests, and time-to-event analyses used Kaplan-Meier curves and Cox proportional hazards. Cluster analysis using k-means was performed on 71 participants with native liver.

**Key Results:** Patients with ALF were significantly younger at enrollment (median 0.8 years vs 3.4 years, p<0.001) and had fewer neurodevelopmental delays (40.0% vs 81.3%, p<0.001) compared to the chronic group. Ascites was more common in ALF (88.2% vs 6.1%, p<0.001). Comprehensive testing was performed more often in chronic than ALF (90.0% vs 43.2%), but etiology was identified more often in ALF (81.3% vs 61.1%). mtDNA depletion was the most common etiology in both groups (ALF: 77% vs Chronic: 41%). Of the 60 participants who underwent DNA sequencing, 38 (63%) had an established genetic diagnosis. Mutations in POLG (n=10), DGUOK (n=5), MPV17 (n=5), and TRMU (n=6) accounted for the majority of cases. All 6 participants with mtDNA deletions or point mutations presented with chronic phenotype. Liver transplant-free survival was 40% at 2 years overall, with ALF significantly worse than chronic (16% vs 65%, p<0.001). Eighteen participants (21%) underwent transplantation; with 33 patient-years of follow-up after transplant, 3 deaths were reported. Cluster analysis identified two clusters, with cluster 2 containing most patients who, despite comprehensive genetic investigation, had no identifiable genetic defect. Higher aminotransferase levels predicted worse outcomes in chronic disease (AST log2 HR 2.51, 95% CI: 1.44-4.38) but better outcomes in ALF (AST log2 HR 0.73, 95% CI: 0.55-0.98). Lactate:pyruvate ratio predicted death/transplant in ALF (log2 HR 2.83, 95% CI: 1.17-6.82, p=0.020).

**Clinical Implications:** This study demonstrates that ALF and chronic MH phenotypes represent distinct clinical entities with different ages at diagnosis, systemic involvement, genetic etiologies, and outcomes. The high diagnostic yield in ALF (81.3% when tested) supports the need for ultra-rapid sequencing in acute presentations to guide liver transplant decisions. The finding that TRMU and DGUOK deficiencies may have better post-transplant outcomes is clinically actionable. The identification of a cluster of patients meeting MH criteria without identifiable genetic etiology (despite comprehensive testing) underscores the need for further research, including whole genome sequencing, multiomics approaches, and consideration of environmental or epigenetic factors. The study also highlights that mtDNA sequencing should be considered in chronic presentations, as large-scale mtDNA deletions were present in 27% of those with chronic disease and a genetic diagnosis.