Association Between Serum Magnesium and Glycemic Control, Lipid Profile and Diabetic Retinopathy in Type 1 Diabetes

Background: Many studies have shown an association between decreased serum magnesium (Mg) levels and poor glycemic control and dyslipidemia in patients with type 1 diabetes (T1DM). However, few studies have evaluated the association between reduced Mg levels and the diabetes complications in these patients and, in particular with diabetic retinopathy (DR), found divergent results. Aims: To evaluate the status of serum Mg levels in adults with T1DM and to assess the association between Mg levels and glycemic control, lipid profile and prevalence of DR. Methods: Retrospective study of adults with T1DM, with an ophthalmological evaluation and a serum Mg level determination. According to Mg levels, the patients were stratified into two groups: normomagnesemic (1.81–2.60 mg/dl) and hypomagnesemic (≤1.80 mg/dl) patients. Exclusion criteria were: patients on diuretics or proton-pump inhibitors, malabsorption or diarrhea, oral magnesium supplementation in recent past, pregnancy or sepsis. Results: Included 105 patients (56.2% male) with median age of 36.0 (interquartile range 16.0) years and median T1DM duration of 16.0 (12.0) years. Median HbA1c was 7.6 (1.5)% and median Mg levels was 1.96 (0.23) mg/dl. Hypomagnesemia (≤1.80 mg/dl) was detected in 20.0% (n=21) patients and 26.7% (n=28) had DR. Hypomagnesemic patients had higher HbA1c [8.2 (1.6) vs 7.5 (1.3)%, p=0.014]. There was no statistical difference in age, sex, T1DM duration or DR between the groups stratified by Mg levels. Mg levels was negatively and weakly correlated with systolic blood pressure (r=-0.200, p=0.041), HbA1c (r=-0.281, p=0.004) and BMI (r=-0.197, p=0.041). There was no correlation between Mg levels and total cholesterol, LDL-C, HDL-C or triglycerides. In multivariate logistic regression analysis, HbA1c was the only predictor of hypomagnesemia [OR=1.541 (1.027–2.312), p=0.037], after the adjustments for age, T1DM duration and BMI. There was no significant difference between patients with or without DR in relation to Mg levels [1.96 (0.28) vs 1.96 (0.19) mg/dL, p=0.986]. Also, there was no statistically significant association between Mg levels and the severity of DR or T1DM duration. In multivariate analysis, T1DM duration, male and estimated glomerular filtration rate >60mL/min/1.73m2 had independently significant association with DR after adjusting for age, glycemic control, hypertension, dyslipidemia, 25-hydroxyvitamin D and Mg levels [OR=1.194 (1.088–1.310), p=<0.001; OR=6.980 (1.654–29.450), p=0.008, and OR=0.780 (0.008–0.751), p=0.028, respectively]. Discussion: Hypomagnesemia is a common problem in adults with T1DM and it was correlated with glycemic control, although we did not find significant association between Mg levels and lipid profile or prevalence of DR. Future longitudinal studies may elucidate the causality between reduced Mg levels and the prevalence of diabetes complications.