Positive correlation between total serum magnesium and serum lipid levels in the general population is clear: Ionized magnesium remains to be examined

We appreciate the comments made by Dr. Corsonello and co-workers but will now address some of the concerns raised by this group. There were questions regarding the population used and the probability of there being high rates of insulin resistance. The study is population based with the only specific limitations on participation being age, inability to attend for blood collection and dual-energy X-ray absorptiometry (DEXA), and presence of severe metabolic, cardiovascular or endocrine disorders. The clinical and biochemical description of the study population are available in this and other publications from our group [ ]. The definition of ‘healthy’ is relative to specifically identified conditions in our exclusion criteria and conditions within our study population that have been specifically identified for examination, such as type 2 diabetes. As to the prevalence of obesity in the study population, 19.9% were obese based on the definition of WHO [ ] using BMI. There is no consensus on definition of obesity based on DEXA measurements at the present time. The reference values for percent body fat based on DEXA are dependent on gender and age [ ]. The relationship between serum magnesium and insulin resistance was analyzed using several approaches including multivariate linear regression analysis. No significant association was detected between serum magnesium and insulin resistance in the entire study, even when subjects with type 2 diabetes were excluded from the analysis. Furthermore, we have subsequently eliminated all obese subjects (BMI > 30), or those with impaired fasting glucose (serum glucose ≥ 5.6 mmol/L), before data analysis in the study group and reanalyzed the data. In all cases the results were qualitatively similar to those of the original work showing serum total cholesterol as an independent predictor of serum magnesium levels.