The underestimated problem of using serum magnesium measurements to exclude magnesium deficiency in adults; a health warning is needed for "normal" results

Background: A major use of serum magnesium measurements in clinical practice is to identify patients with deficiency. However, numerous studies have shown that magnesium deficiency is common and may be present in over 10% of hospitalized patients, as well as in the general population. An important cause for under diagnosis of deficiency is that serum magnesium, the most commonly used test, can be normal despite negative body stores. This article focuses on the limitations of "normal'' magnesium results and highlights the importance of lifestyle or "modus vivendi'' as a pragmatic means of identifying those individuals potentially at risk for negative body magnesium stores. Methods: Researched peer reviewed articles on magnesium published between 1990 and 2008 in MEDLINE and EMBASE, using database keywords "magnesium, deficiency, diagnosis, treatment and hypomagnesaemia''. Bibliographies of retrieved articles have been searched and followed. We have also performed a manual search of each individual issue in which most of these reports have appeared. Results: In 183 peer reviewed studies published from 1990 to 2008, magnesium deficiency was associated with increased prevalence and risk in 11 major conditions. Similarly, in 68 studies performed over the same period, magnesium deficiency was found to predict adverse events and a decreased risk of pathology was noted when supplementation or treatment was instituted. Conclusions: The perception that "normal'' serum magnesium excludes deficiency is common among clinicians. This perception is probably enforced by the common laboratory practice of highlighting only abnormal results. A health warning is therefore warranted regarding potential misuse of "normal'' serum magnesium because restoration of magnesium stores in deficient patients is simple, tolerable, inexpensive and can be clinically beneficial. Clin Chem Lab Med 2010;48:323-7.