The risk of hypophosphataemia in patients receiving intravenous FCM is increasingly recognised; however, the risk in patients with IBD compared with patients without IBD, and associated predisposing factors, have not been previously reported. This prospective observational study demonstrated a mean phosphate reduction of 42% following FCM, similar in patients with and without IBD, with more than half the patients experiencing moderate to severe hypophosphataemia. Although patients with CD more frequently experienced hypophosphataemia than patients with UC, patients with IBD per se were not at greater risk than patients without. Importantly, neither the severity of inflammation (assessed by circulating or faecal markers) nor baseline vitamin D status predicted risk of hypophosphataemia.
Most cases of FCM-associated hypophosphataemia are asymptomatic. Indeed, our data show that delayed adverse events secondary to FCM have no relationship with hypophosphataemia, with the 6 patients with IBD and 8 patients without IBD who experienced fatigue, arthralgia, myalgia, headache and dyspnoea or dizziness have similar phosphate levels to those participants without these symptoms. Hence, hypophosphatemia is difficult to recognise unless specifically measured in the serum. In the largest randomised clinical trial of patients with IBD who received FCM, mean serum phosphate was noted to fall by 38% from baseline (1.12 ± 0.22 mmol/L) to Week 2 (0.69 ± 0.24 mmol/L).[13] Hypophosphataemia as an adverse event was reported in only 6 of 244 patients in that study, and none in a follow-up maintenance study of 104 patients,[12] presumably due to most cases being asymptomatic.
Though most cases of hypophosphataemia appear transient, a minority of patients have persistent reduction in serum phosphate for up to several months. In our current study, serum phosphate remained lower at Day 28 compared to baseline (mean 0.95 vs 1.10 mmol/L, p = 0.001), and in 3 patients remained below 0.6 mmol/L. In another recent study, 56.9% of 52 patients receiving FCM were noted to have moderate to severe hypophosphataemia (defined in this study as < 0.65 mmol/L) at 2 weeks, with 13.7% of patients continuing to have serum phosphate below this level at 6 weeks, and some for up to 6 months.[17] Retrospective studies have also reported hypophosphatemia for as long as 6 months following intravenous iron.[16, 23]
This persistent reduction in serum phosphate, particularly in the context of repeated iron infusions, may contribute to hypophosphatemic osteomalacia with fractures,[18, 24, 25] which may have delayed clinical recognition due to the non-specific nature of symptoms reported by patients and often normal plain radiography. Given that patients with IBD may have nutritional deficiencies and reduced bone density, they may be particularly susceptible to these complications.[19] Furthermore, iFGF-23 is also known to be upregulated by systemic inflammation, potentially further predisposing to hypophosphataemia.[20] Nonetheless, baseline levels of iFGF-23 were similar in patients with and without IBD in the current study. Though not performed in this study, measurement of bone turnover markers following FCM in a future study may help to stratify risk of osteomalacia in the long-term.
Eliciting risk factors for hypophosphataemia following FCM, especially in patients with IBD is, therefore, crucial, to develop potential preventative strategies. The absence of any association between systemic or intestinal inflammation, or vitamin D components (25-hydroxy vitamin D, 1,25(di-OH)D, free or bioavailable 25-hydroxy vitamin D), with risk of hypophosphataemia, means that correction of these factors may not be the answer.
Interestingly, a significant correlation between DBP and the lowest serum phosphate was noted. DBP is a liver-derived α-globulin structurally similar to albumin, which binds about 85-90% of circulating vitamin D metabolites.[26] DBP may control the availability of vitamin D metabolites, especially 25-hydroxy vitamin D, to tissues by allowing only the small free fraction to passively enter cells through diffusion across cell membranes, or actively via interaction with membrane glycoproteins megalin and cubulin.[27] Higher concentrations of DBP may directly reduce circulating 1,25 dihydroxy vitamin D.[28] In contrast, FGF-23 inhibits cytochrome P27B1, the enzyme which 1-hydroxylates 25-hydroxy vitamin D to 1,25 dihydroxy vitamin D. The potential relationship between DBP, FGF-23, 1,25 dihydroxy vitamin D and serum phosphate warrants further study.
The mechanism of hypophosphataemia following intravenous iron administration, specifically after FCM, has been investigated in numerous studies.[15, 29, 30] Consistent with previous reports, the hormone iFGF-23, which primarily inhibits renal phosphate reabsorption producing phosphate wasting but also reduces circulating 1,25(di-OH)D and thus intestinal phosphate absorption,[15, 29, 30] was demonstrated to have significantly risen by Day 2 in our study. Changes in C-terminal FGF-23 should generally be interpreted with some caution as the assay detects both cFGF-23 fragments and the intact molecule. Given that levels remained relatively stable, it is likely that impaired intracellular degradation of iFGF-23 is the likely explanation for the observed changes in FGF-23.[15] The timing of peak of iFGF-23 is uncertain, with initial studies describing a peak at day 1,[15] and other studies reporting a peak at Day 2.[29, 30] Both days 1 and 2 have not been published in a single study, and hence the precise trajectory of iFGF-23 in the first 2-3 days remains uncertain. The decline by day 7 has been consistently reported previously. [15, 29, 30]
Though patients with and without IBD did not differ in rates of hypophosphataemia, patients with CD had a higher risk of hypophosphataemia than those with UC. Subgroup analysis based upon the location of CD was not possible, since only 2 of 11 patients with CD had isolated colonic disease, with most having ileal or ileocolonic disease. Given that phosphate is absorbed in the small intestine, the difference between patients with CD and UC might be accounted for by a greater susceptibility to malabsorption of phosphate in patients with CD and requires further investigation. It is worth noting, however, that baseline phosphate was similar in patients with CD and UC (p = 0.34).
FCM is one of the most commonly prescribed formulations of intravenous iron worldwide, but emerging studies demonstrate a significantly higher risk of hypophosphataemia following FCM compared to other intravenous iron formulations such as iron dextran, iron isomaltoside, and ferumoxytol.[15, 17, 31] The reasons for this difference are unclear, but may be secondary to a differential effect on cleavage of iFGF23 in osteocytes by differing carbohydrate moieties.[15] Although apparently consistent, such serological effects need to be balanced against the relative clinical safety and the limited sequelae of FCM noted to date, particularly in comparison to other formulations such as iron dextran. Longer studies with more rigorous endpoints will enable a clearer distinction to be made.
The strength of this study lies in the uniform, prospective collection of data in patients with and without IBD, and its ability to clarify rates of hypophosphataemia as well as pertinent risk factors. Nonetheless, it must be acknowledged that fractional urinary phosphate excretion was not measured in patients. Previous studies have shown that an increase in phosphaturia is the primary mechanism for hypophosphataemia following intravenous iron, but whether this effect differs between patients with and without IBD, or is influenced by systemic inflammation, is an area for further study. Secondly, the duration of persistence of hypophosphataemia beyond 4-6 weeks and effect on bone turnover markers in this population remains undetermined, particularly in relation to the presence or type of IBD, and remains an avenue for further investigation.