Given the high level of TSAT, the level of hepcidin at diagnosis was inappropriately low, suggesting suppression by an erythroid regulator. The influence of CKD [
6,
7] and metformin [
8] may explain why the hepcidin levels were comparable to those of HC, but its range is consistent with that found in CKD patients [
7]. A metabolic syndrome may also explain high ferritin levels. But these potential biases were internally controlled by comparing two time points before and after vitamin B12 and hemoglobin normalization, when comorbidities and treatments were comparable. ERFE and GDF15 levels were higher than in HC from our institution and previous studies [
9,
10]. ERFE levels at diagnosis were comparable to those of patients with pyruvate-kinase deficiency but lower than those with β-thalassemia [
9]. The inverse correlation of EPO and ERFE kinetics is consistent with that observed in EPO-treated healthy humans [
3] and CKD mouse [
6] models. The GDF15 level at diagnosis was comparable to that of congenital dyserythropoietic anemia type-1 [
11]. Despite the undetectable vitamin B12 levels and typical cytological abnormalities that define megaloblastic anemia, the mean corpuscular volume (MCV) was unexpectedly not macrocytic [
1]. The patient had no history of iron deficiency, inflammation, or red-blood-cell transfusion. GDF15 was recently shown to stimulate erythroid precursor growth in mice [
12]. Cell-cycle acceleration in erythroid precursors may have hidden macrocytosis, but whether high levels of GDF15 may have affected the MCV remains to be demonstrated. Intramedullary hemolysis, suggested by bilirubin, lactate dehydrogenase, and haptoglobin levels, can also explain iron overload [
2]. However, hemolysis alone cannot explain the higher ERFE and GDF15 levels. In sickle-cell diseases (SCD), in which hemolysis is dominant compared with dyserythropoiesis, the ERFE and GDF15 levels and their correlation with hepcidin levels were not comparable to those observed in β-thalassemia [
13]. Additionally, the hepcidin level was inappropriately low relative to that at M3 when TSAT was normal, reinforcing the hypothesis that an erythroid regulator suppressed hepcidin synthesis.
ERFE suppresses hepcidin synthesis by sequestering bone morphogenetic protein (BMP) receptor ligands [
5]. Its synthesis starts upon EPO stimulation through the JAK2-STAT5 pathway in erythroblasts [
14]. ERFE levels are elevated in human congenital or acquired dyserythropoietic diseases associated with iron overload [
3,
15]. GDF15 was shown to suppress hepcidin synthesis in the context of β-thalassemia [
4] and congenital dyserythropoietic anemia [
11]. GDF15 is produced by erythroblasts, but its mechanisms of synthesis and hepcidin suppression are not fully understood [
4,
10].
In conclusion, we found the kinetics of hepcidin, ERFE, and GDF15 to be comparable to those described in iron overload associated with ineffective erythropoiesis (Fig.
1c). This case broadens the spectrum of iron-overload mechanisms in dyserythropoietic anemias to vitamin B12 deficiency, in which hepcidin levels inversely correlate with those of erythroid regulators, suggesting suppression by ERFE and GDF15.