Twelve children (median age 9.5 years) with acute or subacute non-A/E hepatitis were included. Typical symptoms and significantly elevated liver enzymes (including ALT, AST, bilirubin) were present, and coagulation was moderately impaired. Toxicological screening tests were normal.
Upon admission, all SARS-CoV-2 swab PCRs were negative; however, many children had a history of COVID-19 or serological evidence of past infection. Adenovirus PCR in serum was negative in all children; in three children, adenovirus IgG serology was positive, indicating previous infection.
Single-cell analysis showed significant variation. Based on the total number of immune cells in the tissue, two groups were distinguished: one with a high immune cell count (>1,750 cells/mm²; 8 out of 12) and one with a low immune cell count. In the group with a high immune cell count, CD8 T cells, myeloid cells and plasma cells were significantly increased. The infiltrate clearly differed from that seen in autoimmune hepatitis. Transaminases (ALT/AST) were significantly higher in this group; bilirubin, GGT, INR and LDH showed no significant differences. The higher the number of CD8 T cells, granulocytes and myeloid cells in the tissue, the more severe the inflammation; conversely, a higher density of liver sinusoidal endothelial cells was associated with less inflammation.
In a small paired subgroup, the CD8/CD4 ratio in the liver was usually higher than in the blood; activated CD8 T cells, typical of immune cells that have migrated into the inflamed tissue, were found in the liver. In one child with a high immune cell count, pronounced CD8 activation was also detectable in the blood. These findings support the link between local and systemic immune activation, but should be interpreted with caution due to the small number of cases.
In the damaged, collagen-rich areas of the liver, the same neighbourhood of three cell types was repeatedly found: CD8 T cells, macrophage-like myeloid cells and the endothelial cells of the liver. This triad was typical of the damaged areas and corresponds to the strongest signs of inflammation there. In addition, periportal areas with dense immune cell accumulation were found, particularly frequently in biopsies with a high immune cell count (>1,750 cells/mm²).
SARS-CoV-2 proteins (spike/nucleocapsid) were found in eleven of twelve liver biopsies – mostly where the inflammation was most severe. They were mainly found on cells that carry the virus receptor ACE2 (vascular endothelial cells of the liver, macrophage-like defence cells, hepatocytes). These areas were spatially close to activated CD8 T cells and granulocytes. No viral RNA was detectable in the tissue samples tested. This means: proteins yes, but no evidence of ongoing viral replication in the samples tested.
The tissue was also tested for adenoviruses and AAV2:
Ten of the twelve children were treated with steroids during the acute phase. In all patients for whom follow-up examinations were available, the findings improved over time. Two children suffered a relapse after tapering off or discontinuing steroids. Liver transplantation was not necessary in this case series.
The findings may indicate an immune-mediated consequence after recovery from COVID-19: many CD8 T cells and SARS-CoV-2 proteins were found in the most severely damaged areas of the liver, with no evidence of active viral replication. Adenovirus-typical patterns were absent; however, the significance of these findings is limited by the small number of cases and limited tissue material; furthermore, other causes (e.g. drug-induced liver damage) cannot be ruled out with certainty. Further studies should clarify how such cases can best be handled diagnostically and therapeutically.