A bone to pick with ferric carboxymaltose

Abstract

In this issue of Blood, Wagner et al¹ report that the IV iron formulation ferric carboxymaltose (FCM) increases fracture risk in adult cohorts and elucidate underlying mechanisms of skeletal morbidity in a mouse model.

With the realization that IV iron is the preferred form of replacement in a host of conditions associated with iron deficiency, in which oral iron is either ineffective or harmful, the use of IV iron has burgeoned. Multiple formulations allow a complete or near complete replacement dose in 1 brief visit. These formulations, with complex carbohydrate cores that support the safe and slow release of the elemental iron, are low-molecular-weight iron dextran (LMWID), ferumoxytol, ferric derisomaltose (FDI), and FCM. The latter has the unfortunate characteristic of causing hypophosphatemia in 50% to 70% of those to whom it is prescribed.² More concerning is the often protracted nature of the hypophosphatemia, the extreme difficulty in treating it, and the resulting skeletal abnormalities. This most unusual occurrence is fomented by the inhibition of the degradation of the intact and active naturally occurring phosphaturic hormone fibroblast growth factor 23 (FGF23), by the carbohydrate moiety of carboxymaltose by a mechanism yet to be fully elucidated.

Awareness of FCM-induced hypophosphatemia developed slowly and remains low, as many of the symptoms of hypophosphatemia mimic those of anemia and iron deficiency.³ The FCM-induced hypophosphatemia stimulates parathyroid hormone release further enhanced by FGF23’s inhibition of the hydroxylation of 25-hydroxyvitamin D, resulting in a morbid positive feedback loop. As the intact levels of FGF23 begin to fall, any subsequent administration of FCM results in an anamnestic-like response causing protracted and often refractory hypophosphatemia as the administration of phosphorus further increases parathyroid hormone, leading to a worsening secondary hyperparathyroidism. The untoward events have been described to cause nephrolithiasis, osteomalacia, pathologic fractures, and bone deformities.⁴

Wagner et al report a retrospective analysis of an adult, single-hospital cohort treated with either FCM or FDI, demonstrating a higher risk of fractures after FCM treatment over a 7-year period. This association was validated in over 20 000 patients in the TriNetX US Collaborative Network. To investigate the underlying pathophysiological mechanisms, the authors compared effects of IV injections of FCM or FDI on bone in a mouse model of iron deficiency anemia (IDA). Compared with nonanemic controls, adult mice with IDA developed decreases in volume of trabecular bone, iron at the trabecular bone surface, and expression of ossification-related genes in marrow-free bone. Two days after FCM or FDI injection, FCM produced a greater increase in the iron concentration of epiphyseal/metaphyseal regions of marrow-free femurs.

After 7 days, FCM caused a larger percentage of the trabecular bone surface to stain positive for iron, which was detected in endolysosomal compartments of early osteocytes and osteoblasts (see figure). Additionally, FCM was less effective than FDI in upregulating femur expression of bone-formation genes. After 4 weeks, FCM- and FDI-treated mice showed similar improvement in blood hemoglobin levels, but FCM was less effective in improving trabecular bone structure.

FCM and bone biology. In a mouse model of IDA, FCM treatment raised iron concentration in bone, in which iron was detected in endolysosomal compartments of early osteocytes and osteoblasts. In an in vitro model of osteogenic differentiation, FCM reduced extracellular collagen deposition. FCM also reduced the ability of DMP1 to interact with the αVβ3-integrin receptor in vitro and blunted downstream signaling. Figure created with biorender.com. Finberg KE. (2026) https://biorender.com/no27yjj.

Next, the effects of LMWID, FDI, and FCM were examined in a murine preosteoblast cell line that models osteogenic differentiation and extracellular matrix formation. Compared with other iron formulations, cells exposed to FCM showed higher intracellular iron, and only FCM-reduced extracellular collagen deposition. Additionally, FCM-treated cells had higher expression of dentin matrix protein 1 (Dmp1), an extracellular matrix protein and regulator of bone mineralization that belongs to the small integrin-binding ligand, N-linked glycoprotein family.⁵ Similarly in mice, FCM treatment lowered femoral expression of most collagen genes relative to FDI-treatment after 1 week, and FCM significantly raised femoral Dmp1 expression at this time point.

Mice with IDA, as expected, had plasma intact FGF23 elevation, which persisted 1 week after treatment with FCM. Notably, by this time point, FCM-treated mice had not developed hypophosphatemia, which appears later in mice.⁶ DMP1 is known to act as a critical suppressor of FGF23 production in osteocytes.⁵ Intriguingly, the interaction of DMP1 with the αVβ3-integrin receptor was reduced by FCM but not FDI in vitro, and FCM blunted downstream signaling through the MAPK pathway. Collectively, these findings suggest that the detrimental effects of FCM on bone extend beyond hypophosphatemia.

The findings of Wagner et al raise a number of questions. What are the specific physical properties that cause FCM to preferentially accumulate in bone? How does FCM treatment induce DMP1 expression in bone, and to what extent does impaired integrin signaling contribute to the skeletal consequences of FCM? Notably, because dietary iron restriction is usually insufficient to induce IDA in mice, the authors’ model of IDA was generated through a combination of a low-iron diet and serial phlebotomy. However, phlebotomy itself raises plasma FGF23 levels, likely through effects of erythropoietin,⁷ suggesting that plasma FGF23 kinetics observed in the authors’ mouse model may not fully recapitulate FGF23 kinetics in humans with IDA.

Although the skeletal sequelae of FCM described by Wagner et al in adults are of considerable concern, the potential morbidity of FCM in children is even greater. Unfortunately, no literature exists on skeletal outcomes in pediatrics. A retrospective review of children treated with FCM found that although monitoring was uncommon, 19% of those tested had hypophosphatemia, although no skeletal events were described.⁸ In another, retrospective cohort of patients aged <21 years with vigorous testing pre-FCM and post-FCM infusion, 14% developed hypophosphatemia; however, none were severe or protracted, and none were reported to be symptomatic.⁹ However, the full extent of hypophosphatemia in children is unknown, and a prospective clinical trial evaluating phosphorus levels and long-term clinical outcomes in children and adolescents treated with FCM is needed.

The provocative data from Wagner et al add to the growing body of literature supporting deranged bone and mineral metabolism associated with FCM use. FCM-associated hypophosphatemia, secondary hyperparathyroidism, and its incipient skeletal morbidity may be avoided through use of the other 3 available iron formulations, which have been reported to have no disadvantage in safety, efficacy, or cost in prospective, randomized, and double-blind trials.¹⁰