Elsevier

Journal of Critical Care

Volume 42, December 2017, Pages 147-151
Journal of Critical Care

Clinical Potpourri
Total and ionized magnesium testing in the surgical intensive care unit – Opportunities for improved laboratory and pharmacy utilization,☆☆

https://doi.org/10.1016/j.jcrc.2017.07.026Get rights and content

Highlights

  • Category agreement between total Mg and ionized Mg (iMg) is poor.

  • Only 18% of “low” total Mg values are actually low on iMg measurement.

  • Atrial fibrillation and concomitant metoclopramide predict unnecessary Mg testing.

Abstract

Purpose

Ionized fraction (iMg) is the physiologically active form of magnesium (Mg); total Mg may not accurately reflect iMg status. Erroneously “low” Mg levels may result in unnecessary repetitive testing.

Materials and methods

From 11/2015 to 01/2016, patients ordered for Mg from a pilot ICU also had iMg tested. Weighted kappa statistic was used to assess agreement between Mg categories (low, normal, high). Predictors of unnecessary repeated Mg testing and repletion using data were explored through logistic regression models using GEE techniques to account for repeated measurements in both bivariate and multivariable analyses.

Results

There were 470 Mg/iMg paired measurements from 173 patients. The weighted kappa statistic was 0.35 (95%CI 0.27–0.43) indicating poor agreement in assessment of magnesium status. Of the 34 Mg samples reported as “low”, only 6 (18%) were considered “low” using concurrent iMg testing. In the multivariable models, history of atrial fibrillation (aOR = 1.61, 95%CI 1.16–2.21, p = 0.004) and concomitant metoclopramide (aOR = 1.71, 95%CI 1.03–2.81, p = 0.036) were significant predictors of unnecessary repeat Mg testing.

Conclusions

In the surgical ICU, categorical agreement (low, normal, high) was poor between Mg and iMg. Over 80% of “low” total Mg values are erroneous and may result in unnecessary additional measurements and repletion.

Introduction

Magnesium (Mg) is the second most prevalent intracellular cation and is considered an important cofactor in many enzymatic reactions, including protein synthesis, DNA replication, mRNA transcription, mitochondrial function, and energy production through adenosine triphophatase [1]. Derangements in measured magnesium levels in the critically ill are more common than for any other electrolyte [2] and much attention is paid towards maintaining normal Mg homeostasis in these patients, as hypomagnesemia has been associated with neuromuscular symptoms (weakness, delirium, convulsions, etc.) cardiac arrhythmias (premature contractions, atrial fibrillation, torsades de pointes, etc.), metabolic derangements (refractory hypokalemia, hypocalcemia), and increased mortality [1], [3]. Most intensive care units (ICUs) routinely measure total plasma Mg and have developed protocols for Mg administration.

Nearly 99% of the total body magnesium is confined to the intracellular compartment [1], [4]. Extracellular plasma magnesium exists in three states: bound to protein (particularly albumin, 25% of total plasma magnesium), complexed with anions such as phosphate bicarbonate, and citrate (8%), and in ionized form (65%) [5]. Measurement of intracellular magnesium is not possible in routine clinical practice and total plasma magnesium is poorly reflective of intracellular levels [6]. Within the extracellular compartment, the ionized magnesium (iMg) fraction is acknowledged to be the physiologically active form of Mg [7]. Thus, it follows that clinical decisions should be based on the ionized fraction rather than the total Mg level. However, because the total Mg and iMg concentrations are independent of albumin concentration within the normal range, it is not possible to accurately correct for hypoalbuminemia or calculate an iMg; it must be measured directly [8], [9], [10], [11]. Additionally, the correlation between total magnesium and ionized magnesium has been shown to be poor in critically ill patients [12]. While the measurement of iMg has been commercially available since 1994 [13], routine iMg measurement has not become widespread, in part because the iMg assay is only offered by one device manufacturer in the United States. For this reason, most ICUs continue to measure total Mg levels and provide Mg supplementation based on the total plasma Mg.

The primary purpose of this study was to assess the feasibility of implementing iMg into clinical practice and to estimate the potential impact on utilization and laboratory charges. We hypothesized that a large proportion of “low” total magnesium levels would actually be normal on ionized magnesium testing and that substitution of iMg for Mg would lead to significant decreases in repetitive testing and unnecessary magnesium replacement therapy. A secondary objective of this study was to explore potential reasons for why patients received additional magnesium monitoring and replacement outside the bounds of standard protocol.

Section snippets

Material and methods

This prospective study was approved by our Institutional Review Board. We enrolled patients admitted to the surgical ICU of an academic hospital. The surgical ICU admits post-traumatic and post-surgical patients from the following specialties: general surgery, surgical oncology, hepatobiliary, vascular, thoracic, orthopedic, obstetrics and gynecology, and neurosurgery. Medical patients are occasionally admitted secondary to medical ICU bed shortages. There are two 18-bed surgical ICUs in our

Agreement between Mg and iMg

There were 470 pairs of Mg and iMg measurements from 173 patients in the pilot ICU. The Pearson Correlation Coefficient between Mg and iMg was 0.70 (p < 0.0001), showing a moderate correlation (Fig. 1).

Among 470 pairs, 34 (7%) were classified as low, 364 (78%) were normal (1.8–2.4 mg/dL), and 72 (15%) were high (> 2.4 mg/dL) based on the normal range for Mg, while 19 (4%) were classified as low (< 1.8 mg/dL), 325 (69%) were normal (0.45–0.60 mmol/L), and 126 (27%) were high (> 0.60 mmol/L) based on the

Discussion

Our study has several interesting and important findings. First, the majority (> 80%) of plasma Mg tests which are reported as “low” are, in fact, associated with normal levels of ionized Mg, the biologically active fraction. These erroneously low Mg measurements can result in unnecessary medication administration and repeat testing. Second, hypermagnesemia is more prevalent than hypomagnesemia in our surgical ICUs. Finally, we unexpectedly discovered that despite normal Mg measurements,

Conclusion

In the surgical ICU, > 80% of “low” total plasma magnesium values are erroneous and may result in unnecessary additional measurements and unnecessary magnesium supplementation. Hypermagnesemia is common and excessive repeated testing and supplementation represents an opportunity for improvement.

Conflict of interest statement

On behalf of all authors, the corresponding author states that there is no conflict of interest.

References (17)

There are more references available in the full text version of this article.

Cited by (16)

  • Measuring magnesium – Physiological, clinical and analytical perspectives

    2022, Clinical Biochemistry
    Citation Excerpt :

    Most of the methods described below involve the measurement of total serum magnesium, which itself has several limitations, and is typically performed in a central laboratory. The total serum magnesium measurements are not largely influenced by sex or age in the adult population and can be useful in monitoring acute changes, although analytical requirements can be improved for some methods [1,3,9,15,39–41]. Given that total serum magnesium levels are primarily controlled by magnesium excretion in the urine, analysis of magnesium in the urine provides an alternative to total serum magnesium measurements [3].

  • MAGNESIUM

    2022, Laboratory Screening and Diagnostic Evaluation: An Evidence-Based Approach
View all citing articles on Scopus

Funding: This work was supported by Nova Diagnostics (Waltham, MA, USA). The sponsor provided reagents and testing equipment but did not have any role in the design, execution, data collection, data analysis, or manuscript preparation.

☆☆

Conflicts of interest: none.

View full text