Peripheral cytopenias and bone marrow findings in patients with liver failure: A ten-year retrospective analysis at a high-volume transplant center

Abstract Introduction: Peripheral blood abnormalities are a common occurrence in patients with chronic liver disease. Thrombocytopenia is often the first associated cytopenia, occurring in up to 77% of patients with compensated cirrhosis. Anemia and leukopenia are also frequently seen, and when associated with liver disease correlate with increased mortality. The etiology of cytopenias is multifactorial and includes decreased synthesis of thrombopoietin, splenic sequestration, hemolysis of malformed red blood cells, ineffective hematopoiesis, increased blood loss, bone marrow suppression from chronic inflammation, infections, and alcohol use. For patients with advanced liver failure, orthotopic liver transplantation (OLT) remains the only definitive therapy, however its evaluation process is stringent. Bone marrow biopsies (BMB) are occasionally requested as part of the pre-transplant evaluation, however, the role of a BMB in patients with liver failure appears to be limited and rarely diagnostic of a primary hematologic malignancy. This study investigated the role of the BMB and its overall yield for primary hematologic malignancy in patients with liver failure at a high-volume US-based liver transplant center. Methods Our study retrospectively identified adult patients with liver failure who underwent a bone marrow biopsy between August 2012 and February 2025 at the Ochsner Health System. Patients with prior liver transplant, known hematologic malignancies, or cytopenias attributed to non-hepatic etiologies were excluded. Baseline demographics, clinical, and laboratory data including peripheral blood counts, liver disease severity (Child-Pugh and MELD-Na scores), and BMB data were recorded. BMB data recorded included dysplasia, fibrosis, cytogenetics, and Next Generation Sequencing (NGS), if applicable. Peripheral blood counts were analyzed using the Kruskal-Wallis test followed by Wilcoxon pairwise comparisons for subgroup analysis. Results A total of 147 patients met our inclusion criteria with a mean age of 61. 78% of our patients identified as white, 26% as black, and 7% as other. The most common liver disease etiologies were MASH (36%), Alcoholic Liver Disease (26%), and Hepatitis C (16%). Most patients had cirrhosis (93%) with a median MELD-Na of 14 and a Child-Pugh (CP) distribution of 27% class A, 47% class B, and 26% class C. Median spleen size was 15.3 cm. There were 57 total patients (39%) undergoing OLT evaluation at the time of BMB. Pancytopenia was present in 51% of patients and mean WBC, Hgb, PLT, and MCV were 3.25 x 109/L, 9.3 g/dL, 58 x 109/L, and 94 fL, respectively. WBC counts differed significantly across MELD-Na and CP groups (p=0.0002 and p=0.0005) and subgroup analysis showed significant difference between MELD-Na groups 30 and 20 and 20-30 and 20 (p=0.0006 and p=0.0060) and CP class C vs B (p=0.0024) and A (p=0.0002). Hemoglobin differed significantly with all CP class comparisons (B vs A, C vs B, C vs A; p0.0001), however this association was not true for MELD-Na. Platelet count and MCV did not show significant differences when compared to either liver severity score. Bone marrow analysis showed a mean cellularity of 40%, which was normocellular when compared to our cohort's average age of 61. Dysplasia was noted in 33% of patients, fibrosis in 16% (mostly low grade), and only 8% of patients had abnormal cytogenetics. Among 39 BMB samples who underwent NGS, pathogenic mutations were found in 10 and variances of unknown significance in 9. Malignancies were diagnosed in only 5% of patients (n=8), where myelodysplastic syndrome was the most common diagnosis. Of 57 patients being evaluated for OLT, only one was declined based on their new marrow malignancy diagnosis. Conclusion Though cytopenias are prevalent in patients with advanced liver disease, the diagnostic yield of BMB for primary hematologic malignancy is low and rarely alters management. Routine BMB and use of NGS as part of pre-transplant evaluation is of limited value and should continue to be reserved for patients that harbor additional laboratory or clinical features suggestive of advanced hematologic disease.