Drug use is associated with lower plasma magnesium levels in geriatric outpatients; possible clinical relevance

Hypomagnesemia has been associated with diabetes, cardiovascular disease, and other disorders. Drug use has been suggested as one of the risk factors for low magnesium (Mg) levels. In the elderly population, prone to polypharmacy and inadequate Mg intake, hypomagnesemia might be relevant. Therefore, we aimed to investigate associations between drug use and plasma Mg.

Cross-sectional data of 343 Dutch geriatric outpatients were analysed by Cox and linear regression, while adjusting for covariates. Drug groups were coded according to the Anatomical Therapeutic Chemical classification system; use was compared to non-use. Hypomagnesemia was defined as plasma Mg < 0.75 mmol/l and <0.70 mmol/l.

Prevalence of hypomagnesemia was 22.2% (Mg < 0.75 mmol/l) or 12.2% (Mg < 0.70 mmol/l); 67.6% of the patients used ≥5 medications (polypharmacy). The number of different drugs used was inversely linearly associated with Mg level (beta -0.01; p < 0.01). Fully adjusted Cox regression showed significant associations of polypharmacy with hypomagnesemia (Mg < 0.75 mmol/l) (prevalence ratio (PR) 1.81; 95%CI 1.08-3.14), proton pump inhibitors (PR 1.80; 95%CI 1.20-2.72), and metformin (PR 2.34; 95%CI 1.56-3.50). Moreover, stratified analyses pointed towards associations with calcium supplements (PR 2.26; 95%CI 1.20-4.26), insulins (PR 3.88; 95%CI 2.19-6.86), vitamin K antagonists (PR 2.01; 95%CI 1.05-3.85), statins (PR 2.44; 95%CI 1.31-4.56), and bisphosphonates (PR 2.97; 95%CI 1.65-5.36) in patients <80 years; selective beta blockers (PR 2.01; 95%CI 1.19-3.40) if BMI <27.0 kg/m²; and adrenergic inhalants in male users (PR 3.62; 95%CI 1.73-7.56). Linear regression supported these associations.

As polypharmacy and several medications are associated with hypomagnesemia, Mg merits more attention, particularly in diabetes, cardiovascular disease, and in side-effects of proton pump inhibitors and calcium supplements.