Rheumatology and Therapy · Published 2026-03-25 · DOI 10.1007/s40744-026-00842-5
Abstract Introduction The objective of this study is to determine the prevalence of hypomagnesemia in patients with calcium pyrophosphate (CPP) arthritis and the role of diuretics on inducing hypomagnesemia. Methods This was an observational, cross-sectional, prospectively recruited, case–control study. Cases were confirmed by both chondrocalcinosis in plain X-ray and CPP crystals in synovial fluid. Serum creatinine, magnesium, calcium, phosphate, 25-OH-vitamin D, parathyroid hormone, ferritin and transferrin saturation were measured. Diuretic use, class, and dosage were also recorded. Results A total of 568 patients were included, 298 cases and 270 controls. Lower serum magnesium levels were found more frequently among cases, 16% with hypomagnesemia, than in controls [6.4%, (p = 0.01)]. Cases received almost twice as many diuretics as controls (p < 0.001). Comparing cases and controls on diuretic treatment, there were no differences in either magnesium or prevalence of hypomagnesemia. To evaluate the impact of diuretics on serum magnesium, we performed an overall analysis using the entire population. Patients on diuretics showed lower serum magnesium levels compared to patients not on diuretics (1.97 vs 2.05 mg/dl, p < 0.001). Thiazides were associated with lower serum magnesium levels (1.86 vs 2.05 mg/dl; p < 0.001) and hypomagnesemia than loop diuretics (32.1% vs 14.9%; p < 0.004), respectively. Thiazide doses over 12.5 mg/day were associated with higher rate of hypomagnesemia (70.0% vs 11.1%, p = 0.035). Conclusion Patients with CPP crystal deposition disease (CPPD) showed lower magnesium levels and higher rate of hypomagnesemia than controls, and were prescribed almost twice as many diuretics. Both classes of diuretics were associated with lower magnesium levels and clinical significant hypomagnesemia. Thiazides induced hypomagnesemia more frequently than loop diuretics, showing dose-dependent association.
Abstract from DOAJ. Public domain (CC0 1.0).
Read the article at the publisher →