Chapter 3 66 intracellular concentrations of magnesium are associated with glucose and insulin metabolism. The observed positive associations of cholesterol and triglycerides with IEM and plasma magnesium remain difficult to explain. Recently,Waanders et al. studied determinants of hypomagnesemia in diabetic patients and found that higher LDL cholesterol concentrations, but not triglycerides, were associated with higher plasma magnesium concentrations 33. Furthermore, serum magnesium was also positively associated with lipoproteins in a large cohort study including healthy individuals 35. The authors speculated that the positive association might be explained by a binding interaction between serum magnesium and lipoprotein particles. An eGFR below 90 ml/min/1.73m2 was associated with increased plasma magnesium concentrations as well as with decreased 24-h urinary magnesium excretion. Indeed, it is known that in patients with end stage renal disease, plasma magnesium concentration may slightly increase as a consequence of a reduced glomerular filtration rate 36. In the present study, eGFR was not associated with IEM. Finally, IEM was positively associated with age in current study and this association remained significant after adjustment for potential confounders. Interestingly, a previous study found the opposite; intra-erythrocyte magnesium was inversely associated with age 37. We found no association between age and plasma magnesium. It should be noted that this study found several differential associations of dIEM and iIEM with their determinants.Most notably, the association with female sex was stronger with iIEM than with dIEM. Similarly, iIEM also associated more strongly with BMI. These findings implicate that, compared to dIEM, iIEM may lead to higher values in females and in obese participants. On the other hand, the association with plasma potassium was stronger with dIEM. Thus, dIEM may yield higher values in participants with higher plasma potassium concentrations. A few limitations should be addressed. The main limitation is that we were unable to demonstrate cause-effect relationships between magnesium concentrations and clinical features. However, low circulating magnesium concentrations and low urinary magnesium excretions at baseline have been previously associated with increased risk of hypertension 38, and ischemic heart disease 39. Second, previous studies suggested that magnesium is involved in both glucose and insulin metabolism. Unfortunately, insulin was not measured in our cohort and therefore, we are unable to draw conclusions regarding associations between insulin and magnesium. In addition, no data on dietary magnesium intake was available. Instead, however, we used 24-h urinary magnesium excretion as a measure of intestinal absorption 40. Finally, our cohort consists predominantly of Caucasians (98.7%), which limits external validity of our results.
RkJQdWJsaXNoZXIy MTk4NDMw