Elevated serum magnesium associated with SGLT2 inhibitor use in type 2 diabetes patients: a meta-analysis of randomised controlled trials

Aims/hypothesis By analysing available evidence from randomised controlled trials (RCTs), we aimed to examine whether and to what extent sodium–glucose cotransporter 2 (SGLT2) inhibitors affect serum electrolyte levels in type 2 diabetes patients. Methods We searched PubMed, EMBASE, Cochrane Central Register of Controlled Trials (CENTRAL) and ClinicalTrials.gov up to 24 May 2016 for published RCTs of SGLT2 inhibitors that reported changes in serum electrolyte levels. Weighted mean differences (WMD) between each SGLT2 inhibitor and placebo were calculated using a random-effects model. Dose-dependent relationships for each SGLT2 inhibitor were evaluated using meta-regression analysis. Results Eighteen eligible RCTs, including 15,309 patients and four SGLT2 inhibitors (canagliflozin, dapagliflozin, empagliflozin and ipragliflozin) were evaluated. In patients without chronic kidney disease, each SGLT2 inhibitor significantly increased serum magnesium levels compared with placebo (canagliflozin: WMD 0.06 mmol/l for 100 mg and 0.09 mmol/l for 300 mg; dapagliflozin: WMD 0.1 mmol/l for 10 mg; empagliflozin: WMD 0.04 mmol/l for 10 mg and 0.07 mmol/l for 25 mg; and ipragliflozin: WMD 0.05 mmol/l for 50 mg). Canagliflozin increased serum magnesium in a linear dose-dependent manner ( p = 0.10). Serum phosphate was significantly increased by dapagliflozin. Serum sodium appeared to significantly differ by SGLT2 inhibitor type. No significant changes in serum calcium and potassium were observed. Findings were robust after including trials involving patients with chronic kidney disease. Conclusions/interpretation SGLT2 inhibitors marginally increased serum magnesium levels in type 2 diabetes patients indicating a drug class effect. Further investigations are required to examine the clinical significance of elevated magnesium levels in individuals with type 2 diabetes.