Magnesium (Mg2+) is the fourth most abundant cation in the human body only after sodium (Na+), potassium (K+), and calcium (Ca2+), and is the second intracellular cation after potassium (K+) (1). The majority (99%) of magnesium is in the intracellular space, with less than 1% found in the serum. Serum magnesium is comprised of three parts: free or ionized magnesium (iMg), protein-bound magnesium (pbMg), and complexed magnesium (cMg) (2). Serum-ionized magnesium (iMg) is physiologically active. Magnesium is essential for human body health for the reason that ionized magnesium is involved in the interaction of more than 300 enzyme reactions and is important for electrolyte homeostasis, membrane stability, cell division, and generation of action potentials (3–6). Hence, recognition of hypomagnesemia in patients of intensive care unit (ICU) may be important.
Hypomagnesemia has been common, but mostly underdiagnosed electrolyte abnormality. Clinical doctors pay more attention to abnormalities of sodium, potassium, and calcium, but overlook magnesium electrolyte abnormality in clinical practice. Previous studies showed that hypomagnesemia is one of the most common electrolyte disturbances in hospitalized patients, especially in the critically ill. The incidence of hypomagnesemia varies from 20% to 65% in patients admitted to ICU (7–9). The pathology of magnesium deficiencies is multifactorial including gastrointestinal disorders, renal loss, renal diseases, drug-induced loss, metabolic acidosis, and other causes (1, 5). Various hypomagnesemia manifestations may be appeared, such as hypokalemia, hypocalcemia, tetany, vertigo, reversible, psychiatric aberrations, dysrhythmias, hypertension, electrocardiographic changes, acute cerebral ischemia, asthma, and so on (5). Hypomagnesemia is significantly associated with increased need for mechanical ventilation, prolonged ICU stay and mechanical ventilation, increased mortality, and sepsis in critically ill patients. However, those studies had limited patient numbers and the results are controversial (10–13). To comprehensively evaluate the evidence relating to these issues, we conducted this study to assess all available publications that assessed the prognosis of hypomagnesemia in patients admitted to the ICU and draw conclusions from these studies.
MATERIALS AND METHODS
We carried out the study following the statement of meta-analysis of observational studies in epidemiology (MOOSE) (14). Two independent investigators (PJ and FX) carried out the standardized search. Any disagreements were resolved to consensus with a third investigator if necessary. We searched Medline, the Cochrane Database of Systematic Reviews (CDSR), EMBASE databases, which include all papers published up until March 2015, using the following search terms: “hypomagnesemia” or “low magnesium” or “magnesium deficiency” or “critically ill” or “intensive care unit.” Trials were not excluded on the basis of language. All eligible studies were retrieved, and their reference lists were checked for additional articles. To ensure a complete review of the available studies, the abstracts of relevant scientific meetings were also examined.
Study selection criteria
We used each study's definition of hypomagnesemia and hypermagnesemia. Specific inclusion criteria were as follows: studies enrolling patients from ICU (medical or surgical); studies that reported the following outcomes: hospital mortality, mechanical ventilator, sepsis. We excluded literature reviews, studies on animals, or cell lines. Trials published solely in abstract form were also excluded.
Data extraction and quality assessment
This systematic review was undertaken according to PRISMA guidelines. Titles and abstracts were independently reviewed by two reviewers (PJ and FX) to determine their potential relevance. Any disagreements were resolved by consensus with a third reviewer when necessary. Two authors (PJ and FX) independently assessed the risk of bias in the individual trials using Newcastle–Ottawa Scale (NOS). Newcastle-Ottawa Scale was used to assess the quality and authenticity of non-randomized included studies such as cohort and case-control studies (15).
We used Review Manager 5.3 (The Nordic Cochrane Centre, Copenhagen, Denmark) and Stata (Version 10.0, Stata Corporation, Tex, USA) for the analysis. The differences between the two groups were calculated as the relative risks (RR) and 95% confidence intervals (CI) for dichotomous outcomes; the weighted mean difference (WMD) and 95% CI for continuous variables. Heterogeneity was assessed by the Cochran Q-statistic and the I2 statistic. When P is lower than 0.10 and I2 value is higher than 50% indicates significant heterogeneity. If significant heterogeneity existed when it was investigated, we would use subgroup analyses and sensitivity analyses to investigate possible differences between the studies. I2 values of less than 50% represented acceptable between-study heterogeneity, and the fixed-effects model was selected. Otherwise, the random-effects model was selected (16, 17). Sensitivity analysis after excluding one study at a time was performed to assess the stability of the results. Publication bias was determined using the funnel plot and assessed by Egger test (18). A P value of <0.05 was considered to be statistically significant.
Characteristics of the studies
The initial search yielded 249 relevant studies, of which 33 studies were excluded for duplicates. We excluded 152 studies based on titles and abstracts, and 64 full-text articles were assessed for eligibility. After reviewing and accessing the 64 studies, we excluded no relative outcomes for various reasons (animal studies, pediatric patient studies, case reports, editorials, reviews, conference articles, and meta-analyses). Finally, 10 articles were included in the meta-analysis (9–13, 19–23). The detailed steps of the study selection process are shown in Figure 1. The main study characteristics are shown in Table 1. A total of 1,752 critically ill patients were included in these studies.
All included studies reported hospital mortality events (9–13, 19–23), three reported needing for ventilator (9, 19, 20), three reported occurrence of sepsis (10, 11, 19), six reported length of ICU stay (9–12, 19, 20), and three reported length of mechanical ventilation (9, 19, 20).
The definition of hypomagnesemia in ICU patients remains inconsistent in the studies and the different criteria were used in different studies. Hypomagnesemia was defined as magnesium less than 0.42–0.75 mmol/L in these studies. Four studies measured magnesium level within the first 24 h of admission in ICU (10, 11, 20, 22) and six studies measured magnesium level on admission to the ICU (9, 12, 13, 19, 21, 23). Almost all studies achieved such a level one time in their ICU stay, but only one study showed that more than one serum magnesium level was available during the initial 24 h, the lowest value was selected (22). The majority of measurements in the ICU were serum total magnesium in nine studies, but serum-ionized magnesium was measured in one study (12). Seven studies showed that the patients who had documented hypomagnesemia before the ICU admission or who were on magnesium supplementation were excluded (9–11, 19–21, 23). However, the detail of patient's selection for hypomagnesemia was unviable or unclear in three studies (12, 13, 22). We summarized the patient's selection for hypomagnesemia in Table 2.
Hypomagnesemia and hospital mortality
Of all the studies that we included (9–13, 19–23). The average hospital mortality in these studies was 24.3%. Hypomagnesemia was associated with a higher risk of death in adult critically ill patients (P <0.00001). Because of the substantial heterogeneity between studies (I2 = 35%), a fixed-effects model was used to pool RR estimates. The pooled RR was 1.76 (95% CI 1.54–2.00) (Fig. 2).
Hypomagnesemia and the occurrence of sepsis
Three studies reported the incidence of sepsis in ICU patients (10, 11, 19). The data showed that hypomagnesemia was associated with greater risk of sepsis in critically ill patients (RR 2.04; 95% CI 1.21–3.42; P = 0.0007). Statistical heterogeneity was not observed (I2 = 50%; P = 0.14) (Fig. 3).
Hypomagnesemia and the need for mechanical ventilation
Three studies reported ICU patients needing for ventilator (9, 19, 20). The results showed that hypomagnesemia was associated with an increasing rate in need for mechanical ventilation in critically ill patients (P < 0.00001). The RR was 1.36 [95% CI 1.21–1.53], with no heterogeneity (I2 = 0%; P = 1.00) (Fig. 4).
Hypomagnesemia and length of ICU stay
Six reported data regarding length of ICU stay (9–12, 19, 20). The data showed that length of ICU stay was also higher in the hypomagnesemia group (P = 0.01), and the RR was 1.85 (95% CI: 0.43–3.26, I2 = 99%) (Fig. 5).
Hypomagnesemia and length of mechanical ventilation
Three studies were included in this group (9, 19, 20). Length of mechanical ventilation was not different (P = 0.15) between the two groups. Because of the substantial heterogeneity between studies (I2 = 99%), a random-effects model was used to pool RR estimates. The pooled RR was 2.25 (95% CI, −0.80–5.31) (Fig. 6).
To assess the robustness of the stability of the estimate of the length of mechanical ventilation and ICU stay, sensitivity analysis was performed after excluding one study at a time. We found that none of the results was significantly altered, indicating that the results were robust.
Publication bias was detected by Begg and Egger tests. Funnel plots of the 10 studies evaluating the effects of hypomagnesemia on hospital mortality appeared to be symmetrical upon visual examination. The data suggested that there was no evidence of publication bias (Begg test, P = 0.531, Eger test, P = 0.957) (Fig. 7).
Previous studies have demonstrated that hypomagnesemia is associated with clinical prognosis in critically ill patients. However, the results are controversial. In this meta-analysis, we found that hypomagnesemia appears associated with greater risk of mortality, sepsis, mechanical ventilation, and the length of ICU stay in patients admitted to ICU. However, the role of magnesium supplementation for improving outcomes in critically ill patients with hypomagnesemia is needed for further study.
Hypomagnesemia is common in critically ill patients. Hypomagnesaemia causes weakness, muscle cramps, cardiac arrhythmia, and increased irritability of the nervous system with tremors, athetosis, jerking, nystagmus, and an extensor plantar reflex. In addition, there may be confusion, disorientation, hallucinations, depression, epileptic fits, hypertension, tachycardia, and tetany. Hypomagnesaemia is associated with several pathologies, ranging from arrhythmias and pre-eclampsia to cerebral ischaemia (24). Low Mg levels rarely occur alone and clinical manifestations of hypomagnesemia may begin insidiously or dramatically sudden. Hypomagnesemia is easily mistaken for potassium deficit, a condition with which it is often associated. As Kingston et al. concluded that hypomagnesemia rarely shows specific signs or symptoms; its diagnosis depends on a high index of suspicion in patients with hypokalemia, especially after its correction, and in patients with unexplained hypocalcemia (24). We found that some studies assessed the relationships between hypomagnesemia and other electrolytes in this meta-analysis. The criteria for hypocalcemia in the studies included in this meta-analysis were measured serum total magnesium or serum-ionized magnesium on admission to the ICU or the initial 24 h. These studies have inclusion and exclusion criteria for enrolling study subjects. We summarized the patient's selection for hypomagnesemia, definition of hypomagnesemia, and the other electrolytes (e.g., Ca, K) in supplementary Table 1, http://links.lww.com/SHK/A486.
The relation between hypomagnesemia and mortality has varied among studies. Patients with hypomagnesemia have a higher mortality rate than patients with normomagnesemia (9–11, 19–22). However, Curiel-García et al. (21) found no difference in hospital mortality between hypomagnesemic and normomagnesemic patients. Chernow et al. (23) similarly reported no difference in mortality between hypomagnesemic and normomagnesemic patients, but patients with severe hypomagnesemia (serum Mg 1.0 mEq/dL or less) had a higher mortality than patients with normomagnesemia. Soliman et al. (12) also reported there was no significant difference in the length of stay or in the mortality rate between two groups. In this meta-analysis, we found that hypomagnesemia was associated with an increasing rate in hospital mortality in adult critically ill patients (RR 1.76; [95% CI 1.54–2.00]; P <0.00001). Although hypomagnesemia may increase the incidence of hospital mortality in the critically ill patients, the reason for this is not clear. Previous studies suggested that patients with hypomagnesemia have greater incidence of electrolyte abnormalities, cardiac arrhythmias, sepsis, and septic shock, which can increase mortality rates in critically ill adult patients (19). Magnesium is also a cofactor that has a relation with insulin releasing and maintaining sensitivity to insulin (25), and magnesium supplementation decreases insulin requirements (26). In a sepsis-induced animal model, Esen et al. (27) found that magnesium could attenuate the increased blood–brain barrier permeability defect and caused a reduction in brain edema. All these factors might contribute to the higher mortality in critically ill patients in the ICU.
Sepsis is a leading cause of death in critically ill patients admitted to the ICU (28). Sepsis is defined as a systemic inflammatory response syndrome, which can stimulate body to release production of proinflammatory molecules and cytokine (29). Magnesium seems to play an important role in sepsis or septic shock. Magnesium modulates immunological functions such as macrophage activation, leukocyte adherence, granulocyte oxidative burst, lymphocyte proliferation, endotoxin binding to monocytes, and enhanced generation of reactive oxygen species (1, 8). Humphrey et al. (7) reported that there was significant relation between magnesium and inflammatory cytokine production. Mg deficiency can increase interleukin-1, tumor necrosis factor-α, interferon-γ, substance P, and calcitonin gene-related peptide (1). Lee et al. (30) reported that magnesium sulfate had anti-inflammatory effects through activating PI3Kβ, PI3Kδ, and PI3Kγ. In a endotoxemia animal model, Lee et al. (31) found that magnesium sulfate mitigates lung inflammatory response, oxidative stress, and acute lung injury. Our findings indicated that hypomagnesemia was associated with an increasing rate in occurrence of sepsis or sepstic shock in adult critically ill patients (RR 2.04, 95% CI [1.21, 3.42]; P = 0.007). Taken together, these findings may explain why hypomagnesemic patients had higher incidence of sepsis.
Hypomagnesemia is commonly associated with neuromuscular manifestations including muscle weakness, positive Chvostek and Trousseau sign, tetany, muscle cramps, muscle fasciculations, and tremor (1). Muscle weakness is one significant factor that causes difficulty in weaning the patient from the ventilator (19). Some studies reported that Mg deficiency led to a complex array of biochemical, electrophysiological, and morphological abnormalities in skeletal muscle (32). Other study reports that magnesium treatment can prevent diabetic complications such as isometric twitch tensions and resting membrane potentials (33). From our meta-analysis, risk for needing mechanical ventilation was higher in patients with hypomagnesemia.
The study by Upala et al. (34) showed that hypomagnesemia was associated with higher mortality, the need of mechanical ventilation, and also the length of ICU stay in patients admitted to ICU. However, as the authors stated, there are some limitations in their study: first, there was significant heterogeneity for pooled analysis; second, only six studies were included in the study to calculate the mortality outcome and they were restricted to populations mainly from Asia (India, Iran, and China) and Mexico, in the absence of populations in the other parts of the world, which might limit its external validity. Third, the study did not show the pooled data of other ancillary outcomes including sepsis and length of mechanical ventilation. The systematic review by Velissaris et al. (35) showed the increased risk of mortality in hypomagnesemia patients and also performed pooled measure from three studies showing the 1.85-fold increased risk of mortality in hypomagnesemia group. However, as the authors stated, this study performed a systematic review of the relevant literature to define current knowledge in this field and did not use meta-analytic techniques, and did not show the pooled data of other ancillary outcomes including length of ICU stay, mechanical ventilation, and sepsis. Our study has several strengths. First, two independent investigators performed data extraction and data analysis, contributing to the accuracy of data in the meta-analysis. Second, we performed a comprehensive search of the MEDLINE, the Cochrane Central Register of Controlled Trials, and EMBASE databases. Third, there was no publication bias in this study assessed by funnel plot, Egger regression test. Fourth, sensitivity analysis was conducted and confirmed by the robust result. Fifth, our meta-analysis added more studies (10 studies) to increase the strength of the mortality prediction, the need of mechanical ventilation, the length of ICU stay, and also provided the data regarding the association between hypomagnesemia and increased sepsis and longer length of mechanical ventilation. Finally, the populations included in our meta-analysis were not only from Asia (India, Iran, and China) and Mexico, but also from the other parts of the world such as the United States, Belgium, and France, which might exert its external validity. Collectively, the studies by Upala et al. and Velissaris et al. found that hypomagnesemia is associated with higher mortality, the need of mechanical ventilation, and also the length of ICU stay in patients admitted to ICU, a finding strengthened by our present study. Furthermore, we also found that hypomagnesemia is associated with increased incidence of sepsis and the length of mechanical ventilation in critically ill patients. However, consistent with the findings in the previous studies (34–35), high-quality studies are urgently needed to confirm the effect of hypomagnesemia on critically ill adult patients.
It should be noted that our meta-analysis also had some limitations. First, patients in the studies of the meta-analysis were in different ICUs, such as surgical or medical ICU, which may not represent all ICU patients. Second, the definition of hypomagnesemia varied among studies. We also summarized the definition of hypomagnesemia in the other studies (Supplementary Table 1, http://links.lww.com/SHK/A486). We found that the definition of hypomagnesemia was defined as magnesium less than 0.48 to 0.75 mmol/L. Thus, the definition of hypomagnesemia in the studies included in this meta-analysis was similar to the one in the other studies cited in the text.
In summary, this meta-analysis found that hypomagnesemia had strong association with increased incidence of hospital mortality, sepsis, length of ICU stay, and mechanical ventilation in critically ill patients in the ICU. Future studies should evaluate whether magnesium supplementation can improve outcomes in critically ill patients.
1. Noronha LJ, Matuschak GM. Magnesium in critical illness: metabolism, assessment, and treatment. Intensive Care Med
2002; 28 6:667–679.
2. Schenck PA. American Society for Veterinary Clinical Pathology. Fractionation of canine serum magnesium. Vet Clin Pathol
2005; 34 2:137–139.
3. Buckley MS, Leblanc JM, Cawley MJ. Electrolyte disturbances associated with commonly prescribed medications in the intensive care unit. Crit Care Med
2010; 38 (6 suppl):S253–264.
4. Mutlu M, Argun M, Kilic E, Saraymen R, Yazar S. Magnesium, zinc and copper status in osteoporotic, osteopenic and normal post-menopausal women. J Int Med Res
2007; 35 5:692–695.
5. Tong GM, Rude RK. Magnesium deficiency in critical illness. J Intensive Care Med
2005; 20 1:3–17.
6. Koch SM, Warters RD, Mehlhorn U. The simultaneous measurement of ionized and total calcium and ionized and total magnesium in intensive care unit patients. J Crit Care
2002; 17 3:203–205.
7. Humphrey S, Kirby R, Rudloff E. Magnesium physiology and clinical therapy in veterinary critical care. J Vet Emerg Crit Care (San Antonio)
2015; 25 2:210–225.
8. Kvarantan T, Bosman RJ, Oudemans-van Straaten HM. Severe hypomagnesaemia in the intensive care unit. Netherlands J Crit Care
2014; 18 5:23–28.
9. Safavi M, Honarmand A. Admission hypomagnesemia—impact on mortality or morbidity in critically ill patients. Middle East J Anaesthesiol
2007; 19 3:645–660.
10. Chen M, Sun R, Hu B. The influence of serum magnesium level on the prognosis of critically ill patients. Chin Crit Care Med
2015; 27 3:213–217.
11. Zafar MS, Wani JI, Karim R, Mir MM, Koul PA. Significance of serum magnesium levels in critically ill-patients. Int J Appl Basic Med Res
2014; 4 1:34–37.
12. Soliman HM, Mercan D, Lobo SSM, Melot C, Vincent JL. Development of ionized hypomagnesemia is associated with higher mortality rates. Crit Care Med
2003; 31 4:1082–1087.
13. Guerin C, Cousin C, Mignot F, Manchon M, Fournier G. Serum and erythrocyte magnesium in critically ill patients. Intensive Care Med
1996; 22 8:724–727.
14. Rajasekhar A, Lottenberg R, Lottenberg L, Liu H, Ang D. Pulmonary embolism prophylaxis with inferior vena cava filters in trauma patients: a systematic review using the meta-analysis of observational studies in epidemiology (MOOSE) guidelines. J Thromb Thrombolysis
2011; 32 1:40–46.
15. Wells GA, Shea B, O’Connell D, Peterson J, Welch V, Losos M, Tugwell P. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta- analyses. Available at: www.ohri.ca/programs/clinical_epidemiology/oxford.asp
. Accessed March 15, 2015.
16. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ
2003; 327 7414:557–560.
17. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials
1986; 7 3:177–188.
18. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ
1997; 315 7109:629–634.
19. Limaye CS, Londhey VA, Nadkart MY, Borges NE. Hypomagnesemia in critically ill medical patients. J Assoc Physicians India
20. Mousavi SAJ, Salimi S, Rezai M. Serum magnesium level impact on the outcome of patients admitted to the intensive care unit. Tanaffos
2010; 9 4:28–53.
21. Curiel-Garcia JA, Rodriguez-Moran M, Guerrero-Romero F. Hypomagnesemia and mortality in patients with type 2 diabetes. Magnes Res
2008; 21 3:163–166.
22. Rubeiz GJ, Thill-Baharozian M, Hardie D, Carlson RW. Association of hypomagnesemia and mortality in acutely ill medical patients. Crit Care Med
1993; 21 2:203–209.
23. Chernow B, Bamberger S, Stoiko M, Vadnais M, Mills S, Hoellerich V, Warshaw AL. Hypomagnesemia in patients in postoperative intensive care. Chest
1989; 95 2:391–397.
24. Ilicki J, Hardcastle TC. Dysmagnesaemia and outcome in a trauma ICU. S Afr J Crit Care
2014; 30 2:45–50.
25. Huerta MG, Roemmich JN, Kington ML, Bovbjerg VE, Weltman AL, Holmes VF, Patrie JT, Rogol AD, Nadler JL. Magnesium deficiency is associated with insulin resistance in obese children. Diabetes Care
2005; 28 5:1175–1181.
26. Rodriguez-Moran M, Guerrero-Romero F. Oral magnesium supplementation improves insulin sensitivity and metabolic control in type 2 diabetic subjects: a randomized double-blind controlled trial. Diabetes Care
2003; 26 4:1147–1152.
27. Esen F, Erdem T, Aktan D, Orhan M, Kaya M, Eraksoy H, Cakar N, Telci L. Effect of magnesium sulfate administration on blood-brain barrier in a rat model of intraperitoneal sepsis
: a randomized controlled experimental study. Crit Care
2005; 9 1:R18–23.
28. Blanco J, Muriel-Bombin A, Sagredo V, Taboada F, Gandia F, Tamayo L, Collado J, Garcia-Labattut A, Carriedo D, Valledor M, et al. Incidence, organ dysfunction and mortality in severe sepsis
: a Spanish multicentre study. Crit Care
2008; 12 6:R158.
29. Calisto KL, Camacho AC, Mittestainer FC, Carvalho BM, Guadagnini D, Carvalheira JB, Saad MJ. Diacerhein attenuates the inflammatory response and improves survival in a model of severe sepsis
. Crit Care
2012; 16 4:R158.
30. Lee CY, Jan WC, Tsai PS, Huang CJ. Magnesium sulfate mitigates acute lung injury in endotoxemia rats. J Trauma
2011; 70 5:1177–1185.
31. Lee PY, Yang CH, Kao MC, Su NY, Tsai PS, Huang CJ. Phosphoinositide 3-kinase beta, phosphoinositide 3-kinase delta, and phosphoinositide 3-kinase gamma mediate the anti-inflammatory effects of magnesium sulfate. J Surg Res
2015; 197 2:390–397.
32. Okazaki H, Ishimura E, Okuno S, Norimine K, Yamakawa K, Yamakawa T, Shoji S, Nishizawa Y, Inaba M. Significant positive relationship between serum magnesium and muscle quality in maintenance hemodialysis patients. Magnes Res
2013; 26 4:182–187.
33. Pelit A, Emre M, Dagli K, Tuli A. The impact of magnesium on isometric twitch parameters and resting membrane potential of the skeletal muscle in diabetic rats. Cell Biochem Biophys
2013; 65 3:315–319.
34. Upala S, Jaruvongvanich V, Wijarnpreecha K, Sanguankeo A. Hypomagnesemia and mortality in patients admitted to intensive care unit: a systematic review and meta-analysis. QJM
2016; 109 7:453–459.
35. Velissaris D, Karamouzos V, Pierrakos C, Aretha D, Karanikolas M. Hypomagnesemia in critically iii sepsis
patients. J Clin Med Res
2015; 7 12:911–918.