Aplastic anemia (AA) develops in up to 28% percent of children with indeterminate acute liver failure (ALF), often within two months of liver disease onset. Immune destruction of hematopoietic stem cells (HSCs) is thought to underlie its development. We describe a previously healthy 15-year-old female who developed indeterminate ALF for which she received a deceased donor liver graft. Within 2 wks post-liver transplant (pLT), she developed severe cytopenias. A bone marrow (BM) evaluation performed on post-op day (POD)17 revealed 5-10% cellularity, rendering a diagnosis of severe AA (SAA). Given parental hesitation, SAA-directed treatment was held. Filgrastim was initiated on POD19, with her absolute neutrophil count (ANC) rising from 60/µL to > 2,000/µL within 5 d, and discontinued on POD40. Similarly, her hemoglobin began rising on POD35, reaching normal range on POD58. In contrast, her platelet count (plt ct) increased only slowly, peaking at 90,000/µL four months (mo) pLT.

Clues for donor chimerism arose early: The cytogenomic array (CMA) obtained on the POD17 BM failed because the sample appeared to be of mixed origin. While the BM NGS panel lacked Tier I or II variants, eight Tier III variants were reported: four with variant allele frequencies (VAFs) of 0.13-0.14 and four with VAFs of 0.22-0.35. Lastly, the patient's HLA typing, which was performed on blood collected on POD23, could not be reported due to the presence of liver donor-derived DNA.

Short tandem repeat (STR) testing was first performed 5 mo pLT (on lymphocytes only). It showed 55% donor T-cells (B cells were insufficient). Serial testing demonstrated donor chimerism >95% at 16 and 19 mo pLT for T and B cells, respectively. STR testing of monocytes and granulocytes was first performed at 6 ½ mo pLT, showing >95% donor chimerism. All cell populations were 100% donor at 21 and 26 m pLT.

HLA typing of the liver donor and our patient demonstrated 0/10 HLA match, which would be expected to present a high risk of GvHD. Rashes developed at three points pLT, none typical for GvHD. A skin biopsy at 9 ½ mo demonstrated focal vacuolar interface dermatitis, although suggestive of GvHD, was noted to be subtle. The rash resolved quickly without any treatment. At 10 and 13 mo pLT, she underwent upper endoscopies and colonoscopies for abdominal pain and watery stools, with biopsies negative for GvHD. Thus, despite the complete HLA mismatch, she never developed definitive evidence of GvHD.

The BM was monitored serially between 5 ½ and 26 mo pLT, revealing increasing cellularity to 40% and trilineage hematopoiesis. From 5 ½ mo onward, the NGS panel reported just five of POD17's Tier III VUSs. These had the lowest VAFs previously and had increased to VAFs expected for germline (0.46-0.49). Similarly, the CMA, previously uninterpretable, demonstrated focal gains of 19q13.2 and 20q13.33 of uncertain significance at 5 ½ mo and on later testing.

At 7 mo pLT, she was started on romiplostim for recurrent severe thrombocytopenia (nadir 16,000/µL). Her plt ct remained low through the first six wks of increasing doses, then gradually increased to 200,000/µL by ~11 mo pT. Romiplostim was held for periods of 2–4 mo (e.g., 20–24 mo pLT); however, the plt cts progressively dropped to the 140,000s and responded to intermittent low-dose administration (3 mcg/kg).

Immunologic analyses showed progressive increases in T and NK cell % and numbers, achieving normal levels 2 yrs pLT, polyclonal T-cells, and recent thymic emigrants. In contrast, B-cell deficiency persisted. Monthly IVIG was initiated 8 mo pLT for low total IgG levels and continued to maintain trough total IgG ~1000 mg/dL.

Complete liver donor-derived hematopoiesis should be considered in children with HA-SAA pLT who had hematopoiesis recovery without SAA-directed therapy, despite complete HLA mismatch or absence of GvHD. Recipient T cells at the time of LT are likely major determinants of this outcome. Our patient's lymphocyte count was 240/µL on the day of LT and her T cell count was 88/µL when first measured ~ 3 wks later. This case and two others previously reported occurred in children. Although this may reflect the incidence of HA-SAA in children versus adults, other age-related factors may contribute, e.g., number of donor-derived HSCs transferred and size of the recipient. Additional factors may include conditions with extramedullary hematopoiesis or increased circulating CD34+ cells at time of liver donation.

This content is only available as a PDF.
2025
Sign in via your Institution