Hemolysis and Acquired Pyruvate Kinase Deficiency in a Child With a Malignant Myeloid Disorder

Pyruvate kinase deficiency (PKD) is an autosomal recessive disorder characterized by chronic hemolytic anemia. It results as a consequence of homozygous or double heterozygous pathologic mutations of the PKLR gene. Non-genetic or acquired PKD is rare, but has been reported in adults with a variety of blood disorders, particularly in association with myeloid leukemia or myelodysplastic syndrome (MDS) [1]. Usually, this is a clinically insignificant finding, not associated with hemolysis. However, in this letter, we report a child with acute myeloid leukemia (AML) who initially presented with hemolytic anemia and associated PKD.

A previously healthy 8-year-old Chinese boy presented to the emergency department with vomiting, a month of worsening jaundice, and tea-colored urine. He had a several months history of intermittent vomiting and fevers, occurring as often as every 2 weeks, attributed to frequent viral illnesses. His laboratory studies revealed a severe macrocytic anemia with evidence of hemolysis (Table 1A). The direct antiglobulin test (DAT) was positive for C3 and negative for IgG. He was admitted to the hospital for presumed cold agglutinin disease or paroxysmal cold hemoglobinuria. Extensive infectious workup included SAR-COV-2, Flu A/B, HIV, Mycoplasma, CMV, EBV, VZV, and Syphilis, which was negative. His exam was notable for jaundice and scleral icterus, but no splenomegaly was noted by exam or by abdominal ultrasound. Peripheral smear review did not show any specific RBC morphology that would allow characterization of hemolytic anemia. He was transfused with warmed packed red blood cells (PRBC). Repeat DAT was negative on hospital day 4, and his Donath Landsteiner study was negative. Workup for a functional enzyme or membrane disorder was not attempted because he had been transfused with PRBC. However, a next-generation sequencing panel for 51 red blood cell genes (including PKLR) was negative. Despite a good response to red blood cell transfusions, he had a persistence of excess nucleated RBC (250/100 WBC), and small numbers of immature WBC in the form of myelocytes, metamyelocytes, promyelocytes, and blasts, as well as a worsening thrombocytopenia down to a nadir of 68k. Peripheral blood flow cytometry demonstrated 12% myeloblasts. Bone marrow aspirate and biopsy morphology revealed erythroid hyperplasia (M:E ratio 1:3) with erythroid dysplasia (multinuclearity and irregular nuclear budding). Myeloid elements were decreased with a left shift, but no significant dysplasia. Megakaryocytes manifested frequent small size with hypolobulated forms. Bone marrow aspirate flow cytometry revealed increased myeloid blasts (19.9%), but this was considered an overestimation secondary to erythroid hyperplasia from concurrent hemolytic anemia. This was further supported by only mildly increased blasts by morphology and CD34 immunohistochemical stain (3%–4%). Iron stain did not reveal ringed sideroblasts. No cytogenetic abnormalities were detected. FISH studies using MDS probes were negative for Del(5q), Monosomy 5, Del(7q), Monosomy 7, Trisomy 8, Del(20q), and KMT2A rearrangement. Next-generation sequencing was negative for myeloid neoplasm-associated pathogenic mutations, including FLT3, IDH1, IDH2, NPM1, and TP53. Based on these results, the pathology report did not support the diagnosis of a myeloid neoplasia. Following the PRBC transfusion, hemolytic markers improved. The patient was discharged on hospital day 7 in stable condition.

Four weeks following presentation, hemolysis improved somewhat. His DAT was repeated four additional times and all were negative. Platelet count and absolute neutrophil count were normal except for the persistence of 8% peripheral blasts (Table 1B).

Eight weeks post PRBC transfusion, significant hemolysis again was present (Table 1C). An extensive evaluation for hemolytic anemia at this time was negative except for low pyruvate kinase activity compared to other age-dependent enzymes, thus suggesting a pyruvate kinase deficiency. This would represent an acquired PKD because no PKLR mutations previously were detected by next generation sequencing. Again, there were increased peripheral blasts to 12%. Repeat peripheral blood flow cytometry also confirmed 12% myeloid blast but now with slight immunophenotypic aberrancies including brighter than usual CD123 and more uniform than usual CD38. These findings raised a concern for an evolving myeloid malignancy. Simultaneously, next generation sequencing of peripheral blood now showed FLT3 internal tandem duplication. Three months after initial presentation, the patient underwent repeat bone marrow aspirate and biopsy procedure from which 21.5% myeloblasts were noted, consistent with acute myeloid leukemia, which was confirmed by flow cytometry studies.

Acquired red cell enzyme deficiencies are uncommonly associated with dyserythropoietic disorders, myeloid leukemia, and MDS in adults [1-3]. Among acquired RBC enzyme abnormalities, PK deficiency is most frequent. The largest study of acquired red cell enzyme abnormalities in 200 patients, published in 1975, included 72 with leukemia or pre-leukemia, of which 26 (36%) had decreased PK activity [4]. This group manifested 50%–70% of normal PK enzyme activity, similar to what is seen in individuals heterozygous for the PKLR mutation. Also, just like patients with heterozygous mutations, these leukemia/pre-leukemia patients had no evidence of hemolytic anemia.

The clinical course of our patient is most consistent with MDS evolving into AML. He did not meet criteria for refractory anemia of childhood, due to the absence of persistent cytopenias. He met criteria for MDS with excess blasts, although no germline or somatic mutations were identified. Clinically significant hemolysis as a presentation of MDS is highly unusual in children. However, it has been recognized as a presenting symptom of MDS in adults, although poorly characterized. A recent study found approximately 10% of adult MDS patients have non-immune hemolysis [5]. In these cases, the mechanism for non-immune hemolysis has not been defined; however, 2 MDS mutations, U2AF1 and EZH2, have been reported to be associated. Since U2AF1 is involved in RNA splicing in erythropoiesis, it has been hypothesized that an aberrant clone of erythroid precursors with altered RBC biochemical properties may be the cause of the observed clinical hemolysis. Testing of our patient did not reveal mutations in U2AF1 or EZH2.

In one published case report, a 27-year-old male with pre-acute myeloid leukemia presented with hemolytic anemia and acquired PKD [6]. However, to the best of our knowledge, clinically significant hemolysis with acquired PKD and myeloid disease has not previously been reported in children. Our case highlights the need to consider the possibility of underlying myeloid malignancy in children who present with hemolytic anemia, apparent PKD, and other hematologic abnormalities.

The authors have nothing to report.

The authors have nothing to report.

Wendy B. Wong: Nothing to disclose. Michael Jeng: Nothing to disclose. Louise Lo: Nothing to disclose. Bertil Glader: Consultant for Agios Pharmaceuticals.