Magnesium and Electrolytes in Head Injury Cases


The hospital ethics committee approved the study. We measured plasma levels of Mg, P, K, Ca, and Na at admission in 18 consecutive patients with severe head injury (Glasgow Coma Scale [GCS] score, 12) admitted to our ICU (group 1). As controls, we used 19 trauma patients with two or more bone fractures, but no or only mild head injury (GCS 14-15, group 2). Of 19 controls, nine were admitted to the ICU and ten to the high-care unit adjacent to the ICU. We also assessed various clinical variables such as clinical outcome, duration of stay in the ICU, and Acute Physiology and Chronic Health Evaluation (APACHE) II scores at admission in both groups.


Student’s t-test for unpaired results and, where necessary, Fisher’s exact test were used for comparison between groups 1 and 2.


Mean age in group 1 was 42.3 ? 28.1 (range, 17-86) yrs. The average GCS at admission to our hospital was 6.2 ? 4.6 (range, 3-12). Two patients in group 1 used medication that can be associated with loss of Mg and/or P (loop diuretic and Thiazide diuretic, respectively). No preexisting risk factors for electrolyte loss were present in the other patients in group 1. The average age in group 2 was 41.7 ? 20.1 yrs (range, 19-77). None of the patients in group 2 used medication associated with electrolyte disorders.

Electrolyte levels at admission in group 1 vs. group 2 were as follows (mean ? SD): Mg, 0.57 ? 0.17 (range, 0.24-0.85) vs. 0.88 ? 0.21 (range, 0.66-1.42 mmol/L;p <.01). P, 0.56 ? 0.17 (range, 0.20-0.92) vs. 1.11 ? 0.15 (range, 0.88-1.44 mmol/L; p <.01). In group 1, 12/18 patients had Mg levels <0.70 mmol/L vs. 2/19 patients in group 2 ( p <.01); in group 1, 11/18 patients had P levels <0.60 mmol vs. 0/19 patients in group 2 ( p <.01). Nine patients in group 1 had received a single dose of mannitol (50-100 mg) before assessment of electrolytes The average time interval between mannitol administration and electrolyte assessment was <60 mins.

K levels were also lower in group 1 than in group 2 (3.54 ? 0.59 [range, 2.4-4.8[ vs. 4.07 ? 0.45 [range, 3.6-4.8] mmol/L; p <.02. Moderate hypokalemia (K levels below 3.6 mmol/L) was present in 8/18 patients in group 1 vs. 1/19 patients in group 2 (p <.01). Severe hypokalemia (K levels equal or lower than 3.0) was present in 4/18 patients in group 1 vs. 0/19 patients in group 2 (p <.05). Na levels were 146.2 ? 9.1 vs. 138.1 ? 5.8 in groups 1 and 2, respectively. Six of 18 patients in group 1 had Na levels of 150 mmol/L or higher vs. 0/19 in group 2 (p <.05). Ca levels were 2.02 ? 0.24 (range, 1.45-2.51) vs. 2.14 ? 0.20 (range, 1.88-2.46) for groups 1 and group 2, respectively (p =.1).

Fluid resuscitation in group 1 consisted of infusion of saline (NaCl, 0.9%) in 15 patients and of Na 0.45%/glucose 2.5% in three patients Average volume infused before ICU admission was 1060 mL. Two patients had also received blood transfusions. Of the patients in group 2, 12 received infusion of saline (NaCl, 0.9%) and seven received Na 0.45%/glucose 2.5%. Average volume infused before ICU admission was 860 mL. The difference in volume infused between groups 1 and 2 was not significant. Four patients also received blood transfusions before ICU admission. No hypertonic saline was used in our head injury patients.

Urine production in group 1 before admission was measured using a Foley catheter. The average residual urine volume upon insertion of the catheter was 1260 mL in group 1 vs. 380 mL in group 2 (p <.01). The average urine production after catheter insertion but before ICU admission was 360 mL/hr in group 1 vs. 110 mL/hr in group 2 (p <.01). Average urine output in the first 3 hrs of ICU admission was 410 mL/hr in group 1 vs. 90 mL/hr in group 2. This difference was also statistically significant (p <.01). Urine excretion of Na in the first 3 hrs after ICU admission was higher in group 1 (34 vs. 16 mmol/L for groups 1 and 2, respectively).

APACHE II scores were significantly higher in group 1 than in group 2 (22.4 ? 7.8 vs. 6.1 ? 2.1), reflecting differences in GCS as well as other factors, such as tachycardia and tachycardic arrhythmias, episodes of low or high blood pressure, and electrolyte disorders (high Na levels and low K) present in group 1. There were no differences in the presence of chronic diseases between groups 1 and 2. In-hospital mortality was significantly higher in group 1 (72% vs. 10.5%;p <.01). Treatment with one or more antiarrhythmic agents in the course of their stay in the ICU was initiated in 15/18 patients in group 1 vs. 5/19 patients in group 2 (p <.01).


Our results clearly demonstrate that patients with severe head injury are at a high risk for the development of hypomagnesemia, hypophosphatemia, hypokalemia, and to a lesser extent, hypocalcemia when cerebral injury is present. Hypomagnesemia was associated with hypokalemia in most patients.

As outlined in our introduction, hypomagnesemia and, to a lesser degree, hypophosphatemia are associated with various forms of cardiac arrhythmia. The presence of hypokalemia increases the risk of developing arrhythmias.

Cerebral injury in itself is associated with myocardial damage and electrocardiogram abnormalities, including T-wave changes, a shortened P-R interval, a prolonged Q-T interval, premature ventricular contractions, ventricular ectopy, sinus bradycardia, and ventricular and supraventricular tachycardias [15] . Development of electrolyte disorders as reported in the present study may increase the risk and severity of these complications after brain injury. Alternatively, development of these electrolyte deficiencies may be one of the mechanisms through which neurologic trauma is associated with arrhythmias.

Various factors combine to put patients in the ICU at risk for the development of hypomagnesemia and hypophosphatemia. Causes of hypomagnesemia include protein-calorie malnutrition, intravenous administration of Mg-free fluids and total parenteral nutrition, as well as diarrhea and steatorrhea, short bowel syndrome, bowel fistula, and continuous nasogastric suctioning. Renal causes include Bartter and Gitelman syndromes, postobstructive diuresis, postacute tubular necrosis, renal transplantation, and interstitial nephropathy. Medications that can induce renal Mg wasting include loop and thiazide diuretics, aminoglycosides, amphotericin B, cisplatin, pentamidine, and foscarnet. Mg deficiency is seen frequently in alcoholics and diabetic patients. Many of these factors may be present simultaneously in ICU patients. For example, many ICU patients receive Mg-free intravenous fluids and/or total parenteral nutrition, and acute tubular necrosis often occurs in patients the ICU. Trauma patients are frequently treated with antibiotics, often including aminoglycosides.

Print This Post Print This Post

Pages: 1 2 3