Dapagliflozin-associated correction of acute amphotericin B-induced hypomagnesemia

To the Editor, We previously reported the beneficial effects of sodium−glucose cotransporter-2 (SGLT2) inhibitors in severe hypomagnesemia in patients with and without diabetes, including acute hypomagnesemia associated with platinum-based chemotherapy [1–3]. Hypomagnesemia occurs in up to 75% of patients treated with amphotericin B and is less frequent with lipid formulations than with deoxycholate, likely reflecting comparatively reduced distal tubular toxicity, and may occur with or without acute kidney injury (AKI). Herein, we report the use of dapagliflozin to mitigate refractory hypomagnesemia during ongoing intravenous liposomal amphotericin B therapy. A 65-year-old man with orthotopic liver transplantation for hepatitis B− and hepatitis C−related cirrhosis, maintained on tacrolimus (target trough 6–8 ng/ml) and without prior hypomagnesemia, presented with cryptococcal infection requiring two hospitalizations 1 month apart. During the initial admission, he completed a 14-day course of intravenous liposomal amphotericin B (3 mg/kg/day) with flucytosine (100 mg/kg/day) for cryptococcal pneumonia. He developed severe hypomagnesemia (nadir 1.3 mg/dl; mean 1.64 mg/dl over 14 days) requiring 34 g of intravenous magnesium sulfate (MgSO₄) and 8800 mg of oral magnesium oxide. One month later, he was readmitted with cryptococcal meningitis and retreated with the same regimen. Admission serum magnesium was 2.0 mg/dl. By day 4 of amphotericin B, serum magnesium declined to 1.2 mg/dl despite intravenous MgSO₄ (4 g) and oral magnesium l-lactate (168 mg twice daily). Fractional excretion of magnesium, measured >48 h after the last MgSO₄ dose, was 18.2%, confirming kidney magnesium wasting. Concurrent hypokalemia (nadir 3.2 mEq/l) was well controlled with supplementation. Dapagliflozin 10 mg daily was initiated on hospital day 7. Oral magnesium dosing was unchanged, and no further intravenous MgSO₄ was administered thereafter. Serum magnesium increased to 1.5 mg/dl within 5 days of dapagliflozin initiation and remained stable over the subsequent 48 h (Fig. 1). A later episode of diarrhea due to enteroaggregative Escherichia coli, treated with azithromycin, was associated with recurrent hypomagnesemia. Serum magnesium levels during hospitalization. The dashed line indicates the lower limit of the normal serum magnesium range (1.7 mg/dl). Fractional excretion of magnesium (FEMg) was calculated as 100 × (uMg × sCr)/(0.7 × sMg × uCr), where uMg and uCr represent urinary magnesium and creatinine concentrations measured in a random urine sample, and sMg and sCr represent serum magnesium and creatinine levels, respectively. Conversion factors for units: SCr in mg/dl to μmol/l, mg/dl × 88.4; SMg in mg/dl to mmol/l, mg/dL × 0.41152. BID, twice daily; IV, intravenous; EAEC, enteroaggregative Escherichia coli. Post-hoc analyses of randomized trials have shown that SGLT2 inhibitors increase serum magnesium within 4 weeks of initiation [4], and we previously demonstrated this effect within 2 weeks [1]. With daily laboratory monitoring, this case suggests that magnesium stabilization may occur as early as 5 days. Severe amphotericin-associated hypomagnesemia carries risks of arrhythmia, seizures, and treatment interruption. Intravenous magnesium provides transient correction but is inefficient due to rapid kidney excretion, while high-dose oral supplementation is often limited by gastrointestinal intolerance. SGLT2 inhibition in this context may be repurposed as tubular electrolyte modulators, potentially mitigating kidney magnesium wasting despite ongoing tubular injury. Although the precise mechanism remains incompletely defined, SGLT2 inhibitors may enhance tubular magnesium conservation by attenuating tubular injury (potentially via anti-inflammatory effects) and modulating magnesium transport in the thick ascending limb and distal convoluted tubule. This observation provides proof-of-concept for pharmacologic mitigation of drug-induced tubular toxicity, specifically amphotericin-associated kidney magnesium wasting. SGLT2 inhibitors may also attenuate amphotericin B–associated tubular injury through mechanisms we previously described for cisplatin [5] and may reduce AKI risk, particularly important when flucytosine is used concomitantly, given its toxicity in the setting of kidney dysfunction. This report has limitations. We did not obtain repeat urinary magnesium measurements after SGLT2 inhibitor initiation, limiting direct assessment of changes in kidney magnesium handling. Interpretation is further constrained by a confounding diarrheal illness, limiting the evidence of persistent benefit. As a single-patient observation without rechallenge, causality cannot be established. Nonetheless, the temporal association, stabilization of serum magnesium without further intravenous supplementation, and consistency with prior observations support a potential therapeutic signal. Further prospective studies are needed to define the timing, safety, and efficacy of SGLT2 inhibitors for the prevention and treatment of amphotericin B–associated hypomagnesemia and kidney injury. Informed consent was obtained from the patient described in this report All authors declare that there is no conflict of interest.