Sodium

Hypernatremia ([Na+]> 145)
First, assess volume – if invasive hemodynamic monitoring is available, analyze the filling pressures and cardiac output. If these are not available, look to clinical variables such as 1) weight loss 2) peripheral edema 3) [Na+]urine (< 10 mEq/L suggests decreased ECV) and lastly the “shock variables” such as heart rate, blood pressure, urine output, and mental status.

Hypovolemic hyperntremia (most common in general ICU’s):
Causes: loss of hypotonic body fluids (diuresis, diarrhea, vomiting)
Treatment: rapidly replace volume (NS vs 5% albumin), slowly replace
free water (deficit (L) = [0.6 x IBW] x [(Na current/140 ) – 1]), 50%
acutely and the remainder over 24-36 hours to minimize cerebral
edema (ΔNa < 0.5 mEq/L/hr)
Other: Hypovolemia is the most immediate threat (hypoperfusion) but
may not be apparent because of the hypertonicity. The main risk of
hypertonicity is metabolic encephalopathy (seizures, FND, coma) which
in this setting may have a mortality rate as high as 50% [J Crit Illness 11:
720, 1996]. Volume replacement should should be guided by filling
pressures, cardiac output, or the “shock variables.” After volume has
been corrected, tonicity can be corrected with free water
Normovolemic hypernatremia (usually DI):
Causes: nephrogenic DI is caused by hypokalemia, aminoglycosides,
amphotericin, dye, and ATN
Diagnosis: urine < 200 mOsm/L in central and 200 – 500 in
nephrogenic. To diagnose DI, fluid restrict and watch over the next
few hours – if urine osmolality does not increase by 30 mOsm/L,
diagnose DI. This is a dangerous test and these patients must be
watched closely
Treatment: once diagnosed, administer 5 U vasopressin IV. Again, correct free water using the TBW deficit equation (over the course of 2-3 days) and if the etiology is central, give 2-5 U vasopressin SQ q4-6h and monitor [Na+] closely. If DI is not the problem it is most likely a fluid management problem.

Diabetes Insipidus in the Neurosurgical Patient

Hypervolemic hypernatremia (rare):
Causes: hypertonic saline resuscitation, bicarbonate infusion, etc.
Treatment: Most people will autocorrect but if renal Na excretion is
impaired, try combining furosemide with volume replacement
(replacement volume must be of lower sodium concentration than urine)
– you can’t just diurese because you will lose more free water than
volume (because [Na+]urine (~ 75 mEq/L) < [Na+]plasma)

Hyponatremia ([Na+]< 135)
Seen in 50% of hospitalized patients with neurologic problems [Neurologist 12: 117, 2006], 40% of hospitalized AIDS patients [Am J Med 94: 169, 1993], 4.5% of hospitalized elderly patients and 1% of all post-op patients [AIM 117: 891, 1992]. Mortality rates in hyponatremic patients are 2x that of their normonatremic cohort [J Gen Intern Med 9: 89, 1994]. The major complication is metabolic encephalopathy which can lead to CNS pathology as well as ARDS [Chest 107: 517, 1995]. Additionally, overzealous correction can lead to diffuse demyelinating lesions, pituitary damage, oculomotor nerve palsies [Chest 103: 67, 1993], and CPM [Ann Neurol 27: 61, 1990]

Hyponatremia in Neurosurgical Patients

Hypotonic hyponatremia (majority): check volume and urine sodium
Hypovolemic hyponatremia:
Diagnosis: If [Na+]urine > 20 mEq/L probably renal losses (diuresis,
CSW, adrenal insufficiency), if [Na+]urine < 10 mEq/L probably
extrarenal (diarrhea, vomiting w/resuscitation)
Treatment: In symptomatic patients treat with hypertonic saline, in
asymptomatic patients give isotonic saline
Isovolemic hyponatremia:
Diagnosis: SIADH ([Na+]urine > 20 mEq/L, urine osm > 100
mOsm/kg) vs. psychogenic ([Na+]urine < 10 mEq/L, urine osm <
100 mOsm/kg)
Treatment: In symptomatic patients treat with furosemide and
hypertonic saline, in asymptomatic patients give isotonic saline
Hypervolemic hyponatremia:
Diagnosis: renal failure ([Na+]urine > 20 mEq/L) or heart/hepatic
failure ([Na+]urine < 20 mEq/L)
Treatment: in symptomatic patients treat with furosemide and
judicious use of hypertonic saline, if asx just use Lasix
Isotonic hyponatremia: pseudohyponatremia vs. post-urology or post-
gynecology
Hypertonic hyponatremia: hyperglycemia, mannitol

Speed of recovery is controversial (see below). According to Andrews, patients who develop hyponatremia at > 0.5 mEq/L/hr can most likely be treated at 1-2 mEq/L/hr – it is the chronic cases that should be of particular concern. Still, because of the risks inherent in managing hyponatremia, recommendations are that plasma changes do not exceed 0.5 mEq/L/h and that [Na+]plasma does not exceed 130 mEq/L. Calculating the sodium deficit (TBW x {130 – [Na+]plasma }) may help guide resuscitation.
From a review in the New England Journal – “Most reported cases of osmotic demyelination occurred after rates of correction that exceeded 12 mmol per liter per day were used, but isolated cases occurred after corrections of only 9 to 10 mmol per liter in 24 hours or 19 mmol per liter in 48 hours. After weighing the available evidence and the all-too-real risk of overshooting the mark, we recommend a targeted rate of correction that does not exceed 8 mmol per liter on any day of treatment. Remaining within this target, the initial rate of correction can still be 1 to 2 mmol per liter per hour for several hours in patients with severe symptoms. Should severe symptoms not respond to correction according to the specified target, we suggest that this limit be cautiously exceeded, since the imminent risks of hypotonicity override the potential risk of osmotic demyelination.” Over-rapid correction is rarely necessary as “…even seizures induced by hyponatremia can be stopped by rapid increases in the serum sodium concentration that average only 3 to 7 mmol per liter” [NEJM 342: 1581, 2000] There is a relative lack of evidence regarding rapid correction of acutely-hyponatremic patients, probably because 1) most patients will regain their neurologic baseline after a 3-7 mEq/L correction and 2) CPM is devastating and while the original data was in chronic patients (mostly alcoholics), fear of CPM stamped out rapid correction in known acute cases. There was one study of rapid correction in acutely hyponatremic patients, which showed no complications (6 years later) after correcting on average 21.6 mM/L in 24 hours in acutely symptomatic hyponatremic psychogenic water drinkers [Am J Med. 88:561, 1990]

Hyperglycemia:
These patients are profoundly volume depleted and should be treated
similarly to hypovolemic hypernatremics – first replace the volume deficits,
then the free water. Remember to correct the sodium (1.6 per 100 until 400
then 4.0 per 100 after that) – brain cell volume restoration can be particularly
rapid in these patients, so proceed with caution [9]. Insulin should actually be
held until volume is given as it can drive glucose (and water) into the cells –
thus, don’t give insulin until after volume has been started

Potassium

Only 2% of total body potassium is extracellular (0.004% in plasma), and this severely limits the value of serum potassium measurements – the relationship between serum and total body potassium is non-linear such that at [K+] < 3.5, small decreases in serum levels reflect huge decreases in total body K+, whereas at [K+] > 5.0, large increases in serum levels reflect small increases in total body K+ [Kidney Int 30: 116, 1986]

Hypokalemia (transcellular shifts vs. total body depletion)

β2-agonists are known to cause hypokalemia via transcellular shift, but at therapeutic doses this is usually ~ 0.5 mEq/l [Ann Emerg Med 21: 1337, 1992]. On the other hand, when combined with diuretics the additive effects can be significant [Am J Med 86: 653, 1989]. Other causes of transcellular shift into the cell are alkalosis, insulin, and hypothermia

The leading cause of renally lost hypokalemia is diuretic therapy. In the ICU, additional culprits include NG suctioning, magnesium deficiency [Crit Care Clin 7: 225, 1991], and alkalosis (actually due to shift, see above) – to help differentiate these, analyze the urine chloride (> 25 mEq/L for diuretics or Mg deficiency, < 15 mEq/L for NG drainage or alkalosis). The leading cause of extrarenally lost hypokalemia is diarrhea

Patients usually asymptomatic until 2.5 mEq/L, at which point look for weakness and mental status changes, as well as > 1 mm U waves, loss of T waves, and QT prolongation. Hypokalemia alone does not cause arrhythmias – rather, when combined with another abnormality (such as dig toxicity, magnesium depletion), it can act to enhance them

Management should first focus on reversing transcellular shifts (ex. alkalosis, hypothermia). After these have been ruled out, proceed to potassium replacement. KCl is available in 1.5 and 2 mEq/L concentrated solution (ampules of 10, 20, 30, 40 mEq), but these are 4000 mOsm/L and must be diluted [Trissel LA Handbook Inj Drug: 886, 1994]. KPO4 is also available and some prefer it in DKA because of the potassium depletion. Always correct calcium as well because hypokalemia protects against the adverse effects of hypocalcemia

Standard infusion rate is 20 mEq in 100 mL of isotonic saline over 1 hour, although 40 mEq may be needed if levels < 1.5 mEq/L or serious arrhythmias are present (80 mEq have been given safely [Crit Care Clin 7: 143, 1991]). Up to 20 mEq/hr can be given through a central line (and should be because of venous irritation), but above that the dose should be split between two peripheral lines to avoid the possibility of transient cardiac standstill. Remember that because of the potassium curve, serum levels will be slow to change following hypokalemia – still, if [K+] levels appear refractory, check magnesium because this can be a cause of continued potassium loss [Arch Intern Med 145: 1686, 1985]. ALWAYS PUT PATIENTS ON A CARDIAC MONITOR WHEN GIVING INTRAVENOUS POTASSIUM

Hyperkalemia (transcellular shifts vs. insufficient excretion vs. overload)

The most common cause is artifact, either due to hemolysis (20% of the time [Arch Intern Med 147: 867, 1987]), distal muscle release secondary to the tourniquet [NEJM 322: 1290, 1990], or release during clot formation in the specimen tube. For this reason, a spurious level in an otherwise normal patient should not invoke treatment until a second lab is confirmatory. If real, this is more dangerous than hypokalemia

β-blockers and digitalis can cause hyperkalemia via transcellular shifts, as can lack of insulin ([K+] > 7 mEq/L can only be caused by digitalis toxicity). Besides drugs, myonecrosis can cause massive release of potassium but until renal function deteriorates, K+ should be adequately cleared. Acidosis is often blamed for hyperkalemia, but respiratory acidosis is not always accompanied by hyperkalemia [Hall RI et al. Anesth Analg 76: 680, 1993] and there is no clear evidence that organic acidosis (LA, KA) can produce hyperkalemia [Am J Med 71: 456, 1981; Hall RI et al. Anesth Analg 76: 680, 1993; NEJM 297: 796, 1977] – it may be that the observed hyperkalemia simply accompanies renal failure or RTA

Renal insufficiency can produce hyperkalemia if GFR < 10 mL/min or UOP < 1L/day [J Intensive Care Med 3: 52, 1988]. Other renal losses occur in hypoadrenergic states and drugs such as ACE-I, K-sparing diuretics, and NSAIDS (and to a lesser extent heparin, TMP-SMX, cyclosporine, and succinylcholine). When the source is not clear (cellular shift vs. insufficient losses), get a urine potassium (> 30 mEq/L suggests shift, < 30 mEq/L suggests insufficient loss)
After 14 days, 1 unit of PRBC provides a 3.1 mEq potassium load, and a unit of whole blood provides 4.4 mEq [Transfusion 15: 144, 1975]. This is adequately cleared except in circulatory shock, ie massive resuscitations. Because their VD is reduced, these patients can become hyperkalemic

Cardiac manifestations of hyperkalemia start at 6.0 mEq/L with T waves, elongated PR interval, decreased P wave amplitude, widened QRS (treat rapidly), sine waves, and finally asystole

To treat critical cases, stabilize with Ca2+(see table for dose depending on clinical state). ONLY GIVE CaCl2 IF QRS WIDE (CaCl2 10cc IVP over 10 min except kids or non-emegencies, they get gluconate). For those taking digitalis, give 10% gluconate diluted in 100 mL isotonic saline over 20-30 min and in those with digitalis toxicity, give MgSO4 instead

Acute Management of Hyperkalemia[edit]

Condition	Treatment	Comment
EKG changes and circulatory compromise (or just wide QRS)	CaCl (10%) 10 mL IV over 3 min	For anyone with wide QRS
EKG changes or K > 7 w/o circulatory compromise	CaGluc (10%) 10 mL IV over 3 min repeat after 5 min if needed	Response lasts ~ 25 min, do NOT give bicarbonate after calcium
AV block refractory to Ca2+	10 U regular insulin in 500 mL of 20% dextrose infused over 1 hr/Transvenous pacemaker	Should drop K by 1 mEq/L for 1-2h
Digitalis cardiotoxicity	MgSO4 2 g IV bolus; Digibind	Do not use calcium in digitalis toxicity!!!
After the acute phase has resolved or when the EKG stops changing	Kaeyexalate 30 g in 50 mL of 20% sorbitol PO or 50 g in 200 mL 20% sorbital PR	Oral dosing preferred

Drive in with insulin/dextrose and maybe β-agonists, give volume to dilute. Bicarbonate can be tried if acidotic, but note that most patients will be in renal failure, and in this scenario insulin works much better than bicarbonate [Am J Med 85: 507, 1988] and also bicarbonate binds Ca2+ and should not be given in patients who received it. For these reasons, bicarbonate is rarely useful. Remove with Kayexelate and dialysis. Note that with Kayexalate, every 1 mEq of K lowered will raise Na by 2-3 mEq. Hemodialysis is the most effective method of removal in patients in renal failure [Am J Med 85: 507, 1988]. Lasix, Mg2+, and bicarbonate are rarely used, because these patients are so often in renal failure

Magnesium (All ICU patients get daily Mg unless in renal failure)

Less than 1% of total body magnesium is found in plasma, limiting the value of [Mg2+]serum to estimate true body stores. In fact, patients who are magnesium deficient can have normal serum levels [Arch Intern Med 148: 2415, 1988]. When analyzing [Mg2+], use serum and not plasma because plasma measurements may contaminate with citrate or other anions that bind Mg2+ [Clin Chem 33: 1965, 1987]. The standard assay measures all Mg2+, not just the ionized fraction, which may further complicate analysis. In a normal patient [Mg2+]urine is MUCH more indicative of total body stores than [Mg2+]serum [Medicine 48: 61, 1969], however most cases of depletion involve renal losses in which case its diagnostic value is limited

Magnesium Deficiency

Magnesium deficiency (+/- hypomagnesemia) is found in 15% of ward patients [JAMA 263: 3063, 1990; Crit Care Med 21: 203, 1993] and 60% of ICU patients [Crit Care Med 13: 19, 1985; Chest 95: 391, 1989]. It may be the most underdiagnosed electrolyte abnormality in medicine [Am J Med 82: 24, 1987]

Diuretics are the leading cause of Mg deficiency, particularly the loop diuretics – it may occur in up to 50% of patients on furosemide [Am J Med 82: 11, 1987]. Thiazide diuretics show similar effects but only in the elderly [Am J Med 63: 22G, 1989]. Potassium-sparing diuretics do not cause depletion [Am J Med 82: 38, 1987]. Aminoglycosides (30% of patients), amphotericin, and pentamidine can all cause magnesium depletion [Surg Gynecol Obsstet 158: 561, 1984], as can digitalis and some chemotherapeutic drugs. 30% of patients admitted for EtOH abuse and 85% of those in DTs will have hypomagesemia [Crit Care Clin 5: 217, 1985]. There is an association between magnesium and thiamine deficiency, as Mg is required to convert thiamine into thiamine pyrophosphate [Acta Med Scand 218: 129, 1985] – anyone on thiamine should have Mg monitored. Diarrhea can lead to depleted Mg, as can diabetes – while only 7% of DKA admissions are hypo-Mg, within 12 hours 50% will be, probably because insulin drives it into cells [Lau K: Mg Metabolism; Livingstone 575, 1985]. 80% of patients with acute MI will be hypo-Mg within 48 hours of presenting [Arch Intern Med 147: 753, 1987]

There are no clinical signs specific to hypo-Mg, but it is often associated with other electrolyte abnormalities – it helps cause hypocalemia and hypocalcemia, but is caused by hypophosphatemia (so always replete phosphate if both are low). Mg can magnify the effect of digitalis on the heart, leading to cardiotoxicity – IV Mg is often effective in suppressing digitalis-toxic arrhythmias even when [Mg2+]serum is normal [JAMA 249: 2808, 1983]. It is known for its effects on Torsades but it can sometimes help resolve other refractory arrhythmias even in the absence of Mg depletion [Am J Cardiol 65: 1397, 1990]. Neurologic findings include AMS, seizures, tremors, and hyper-reflexia. “Reactive CNS Mg Deficiency” produces ataxia, slurred speech, metabolic acidosis, excess salivation, seizures, muscle spasms, and obtundation which is often brought on by loud noises or bodily contact [Arch Intern Med 151: 593, 1991]

In the absence of renal losses, urinary excretion in response to a Mg load is the most sensitive index of total body stores [Arch Intern Med 148: 329, 1988; Lancet 340: 124, 1992]. Give 24 mmole of Mg (6 g MgSO4) in 250 mL NS over 1 hour and collect urine for 24 hours. If > 80% of infused Mg is excreted, deficiency is unlikely, whereas if < 50% are excreted, the individual is probably deficient. Remember, however, that this test is unreliable in patients with impaired renal function. In those with normal renal function it is useful as a screening test and also to titrate replacement therapy. Also remember that urine excretion may not be useful in the ICU population because hypo-Mg is often caused by renal losses

PO Mg can be used to treat mild cases but is absorbed erratically, so IV should be used in symptomatic or severe cases. 1 g MgSO4 has 8 mEq (4 mmole) of Mg ions. Because of its high osmolarity, the 50% preparation must be diluted to 20% for IV use. Dilute in NS and not LR as Ca2+ ions will counteract the Mg2+

Replacement Protocols for Magnesium Deficiency

Note that one study of ionized magnesium levels after traumatic brain injury showed a correlation between low [Mg2+]i and degree of injury, but not [Mg2+]total [Scand J Clin La Inv 55: 671, 1995]. Based on this, Andrews recommends considering Mg supplementation even in patients with normal levels

Hypomagensia has been associated with vasospasm in both animal and human models [Stroke 22: 922, 1991]. In one study, 283 SAH patients randomized within 4 days IV MgSO4 vs. saline showed favorable trends in delayed cerebral ischemia (reduced by 34%, 95% CI 0.38–1.14) and in reduction in poor outcome at 3 months (23% risk reduction, risk ratio 0.77, 95% CI 0.54–1.09) [Stroke 36:1011, 2005]. In another study, 60 SAH patients randomized, to MgSO4 80 mmol/day vs. saline infusion for 14 days (both received intravenous nimodipine, hypertensive and hypervolemic therapy) showed a trend towards decreased vasospasm (43% to 23%, p=0.06) and a significant decrease in duration of Doppler-detected vasospasm (mean flow velocity >120 cm/s and a Lindegaard index >3, p<0.01) [J Neurosurg Anesthesiol 142: 142, 2006]

Lastly, low magnesium levels are seen in experimental brain and spinal cord models [J Neurotrauma 15: 183, 1998; J Neurotrauma 10: 215, 1993]

Hypermagnesemia

Hypermagnesemia may be found in 5% of hospitalized patients [JAMA 263: 3063, 1990]. Predisposing conditions include massive hemolysis, renal insufficiency (CrCl < 30 mL/min), DKA (transient), adrenal insufficiency, hyperparathyroidism, and lithium toxicity [Crit Care Clin 7: 215, 1991]. Mg is a calcium channel blocker and thus its prominent effects are cardiac [Am Heart J 108: 188, 1984]

Features of Hypermagnesemia

Level	Clinical Findings
4.0 mEq/L	Hyporeflexia
5.0 mEq/L	Prolonged AV conduction
10 mEq/L	Complete heart block
13 mEq/L	Cardiac arrest

Magnesium in Subarachnoid Hemorrhage Patients

Treatment of Hypermagnesemia

To treat hypermagnesemia, start with IV CaGluconate (1 g over 2-3m) to buy time before dialysis is started. If the patient has normal renal function and fluid status, aggressive fluid resuscitation and furosemide can be added as well

Calcium

Standard calcium assays measure all three forms (protein bound, chelated, and free). Several “corrective factors” are available but none of them are accurate [J Nutr 120S: 1470, 1990] – the only way to accurately measure physiologically-relevant calcium is to measure the ionized fraction directly. Acidosis can lead to elevated ionized calcium. Gas bubbles in the blood sample can expel CO2, falsely lowering the ionized calcium. Lastly, anticoagulants (heparin, citrate, EDTA) can bind calcium, so in general place these blood samples in red-topped tubes. Heparinized tubes can be used if [heparin] < 15 U/mL [Ann Clin Lab Sci 21: 297, 1991]

Normal Ranges of Ca2+ and PO42-
Electrolyte mg/dL SI units (mM)
Total Ca2+ 8.0 – 10.2 2.2 – 2.5
Ionized Ca2+ 4.0 – 4.6 1.0 – 1.5
Phosphorus 2.5 – 5.0 0.8 – 1.6

Hypocalcemia

Ionized hypocalcemia is found in 50 – 65% of ICU patients [Contemp Surg 45: 71, 1994; Crit Care Med 20: 251, 1992] and in 15-20% on admission to the ICU

Causes of Ionized Ca2+ Depletion in the ICU

% = amount of those with given condition who will be hypocalcemic
Mg2+ depletion (70%) Aminoglycosides (40%) Alkalosis (resp.)
Renal insufficiency (50%) Cimetidine (30%) CABG
Sepsis (30%) Heparin (10%) Pancreatitis
Blood transfusion (15%) Theophylline (10%) Fat embolism

Mg2+ depletion inhibits PTH secretion and leads to a hypocalcemic state which is refractory to Ca2+ replacement [thus always replete magnesium in these patients]. Sepsis is a common cause of hypocalcemia but it is not known whether or not low Ca2+ is clinically significant in these patients [Arch Surg 127: 265, 1992]. Alkalosis can cause hypocalcemia, but usually only if respiratory in origin. Blood transfusions contain citrate which binds Ca2+, but this is usually transient as the liver and kidneys rapidly metabolize the citrate [Zaloga GP in Chernow B, ed. W&W p. 777, 1994] – in renal failure, the effect may last significantly longer

Renal failure causes hypocalcemia because of insufficient vitamin D conversion, and traditionally this was treated by lowering serum phosphate levels with antacids, however the value of this is totally unproven

Clinical Effects of Hypocalcemia

There are two major clinical effects of hypocalecmia – neuromuscular excitability and cardiovascular depression. Beware reliance on the clinical picture – Chvostek’s sign (facial nerve) is present in 25% of normal adults, and Trousseau’s sign (arm) is missing in 30% of hypocalcemic patients [Zaloga GP in Chernow B, ed. W&W p. 777, 1994]. Hypotension, decreased CO, and ventricular ectopy can all follow hypocalcemia but usually do not occur until ionized levels drop below 0.8 mM [Zaloga GP in Chernow B, ed. W&W p. 777, 1994]

Intravenous Ca2+ Replacement Therapy

Solution	Elemental Ca2+	Unit	Volume	Osmolality
10% CaCl	27 mg/mL	10 mL	ampules	2000 mOsm/L
10% Ca Gluconate	9 mg/mL	10 mL	ampules	680 mOsm/L

Symptomatic hypocalcemia is a medical emergency and should be treated intravenously but only if [Ca2+] < 0.65 mM (see below). The above IV solutions are hyperosmolar and should be given through a central vein if possible – if a peripheral vein is used, gluconate is preferred. A bolus of 200 mg Ca2+ diluted in 100 mL saline will raise [Ca2+] by 1 mg/dL (or 0.25 mM) but this will begin to drop within 30 minutes, so maintenance of 1-2 mg/kg/h is required for at least 6 hours. After calcium resuscitation, the patient can be maintained on PO calcium (2-4 mg/day). Caution: IV calcium can promote vasoconstriction and/or ischemia in any organ and the risk is much higher in low CO patients who are already maximally vasoconstricted [New Horiz 4: 134, 1995]. Additionally, calcium overload can produce lethal cell injury in patients in circulatory shock [New Horiz 4: 139, 1995]

Hypercalcemia

Rare, found in < 1% of hospitalized patients [Q J Med 77: 1277, 1990]. 90% of cases are caused by hyperparathyroidism or malignancy [Hosp Pract 29: 79, 1994], but more rare cases (ie < 1:1000 hospitalized patients) include immobilization, thyrotoxicosis, and drugs. Clinical manifestations are non-specific and include GI, cardiovascular (shortened QT interval, hypotension/hypovolemia), renal (nephrocalcinosis, polyuria), and neurologic (confusion). Patients require treatment if symptomatic or if ionized Ca2+ > 3.5 mM (14 mg/dL)

Hypercalcemia usually produces hypercalciuria which produces osmotic diuresis – these patients are volume depleted and therefore require NS infusions. Saline infusion alone will not return [Ca2+] to normal levels, so furosemide (40 – 80 mg IV q2h, UOP 100-200 mL/h) is added to enhance excretion. Keep in mind that any additional losses due to diuresis must be replaced with NS. Calcitonin 4U/kg q12h can actually reverse bone resorption, and works within a few hours but can only drop serum calcium by ~ 0.5 mM. Bisphosphonates are more potent anti-resorption agents but take longer to work – pamidronate 90 mg IV infusion over 2 hours usually takes 4 – 5 days to reach peak effect. Zolendronate (4 mg over 15 min) is an alternative. Hydrocortisone 200 mg IV qday, when combined with calcitonin, can be very useful, especially in renal failure or multiple myeloma [Hosp Pract 29: 79, 1994; NEJM 326: 1196, 1992]. Dialysis is a last resort

Treatment of Severe Hypercalcemia

Agent	Dose	Comment
Normal saline	Variable	Initial treatment of choice
Furosemide	40-80 mg IV q2h	Titrate to UOP 100-200mL/hr
Calcitonin	4U/kg IM or SQ q12h	Fast but not very effective
Hydrocortisone	200 mg IV daily in 2-3 divided doses	Excellent adjunct to calcitonin
Pamidronate	90 mg by continuous IV infusion over 24 hours	Delayed response but very potent

Phosphorus

Participates in glycolysis and high energy phosphate production

Normal Range of PO42-
Electrolyte mg/dL SI units (mM)
Phosphorus 2.5 – 5.0 0.8 – 1.6

Hypophosphatemia

Defined as < 0.8 mM (or < 2.7 mg/dL). Occurs in 17-28% of ICU patients [Crit Care Resusc 6: 175, 2004]. When glucose moves into cells, PO42- usually follows it. Not surprisingly, glucose loading is the most common cause of hypophosphatemia in hospitalized patients and occurs often during the re-feeding of alcoholic or otherwise malnourished patients [Ann Pharmacother 28: 626, 1994; Arch Intern Med 148: 153, 1988]. It can occur with PO, enteral, or parenteral feedings and is one reason why parenteral feeds are started gradually (see figure below). Respiratory alkalosis elevates pH which accelerates glycolysis, ultimately moving phosphate into cells and leading to hypophosphatemia – this may be a very important cause in ventilator-dependent patients. Sucralfate is associated with hypophosphatemia presumably due to its binding activity [Ann Emerg Med 21: 1337, 1992]. Similarly, Al(OH)3 antacids bind phosphate and reduce serum levels. PO42- depletion is nearly always found in DKA but correction does not seem to affect outcome. β-agonists and sepsis are associated with hypophosphatemia but the significance of this is unclear

Hypophosphatemia is often clinically silent, and even at levels < 1.0 mg/dL may not produce any obvious effects [South Med J 80: 831, 1987]. Still, in certain patients hypophosphatemia has been shown to be deleterious. Hypophosphatemia can lower cardiac output and patients with low [PO42-] and in heart failure may respond to supplementation [Am J Med Sci 295: 183, 1988]. Hypophosphatemia can lead to deformed RBCs and severe hypophosphatemia can lead to hemolytic anemia [Ann Pharmacother 28: 626, 1994]. It also shifts the O2-Hb curve left, reducing tissue release of O2. It can impair WBC function and has been shown to cause several neurologic manifestations (confusion, seizures, coma). Low levels of PO42- reduce its availability for high energy bond formation. There is one report of hypophosphatemia associated with muscle weakness and respiratory depression, but this appears to be an isolated incidence and is probably insignificant [Am J Med 84: 870, 1988]

IV replacement is recommended for anyone with levels < 1 mg/dL (0.3 mM) or with adverse effects or comorbid disease such as cardiac dysfunction, respiratory failure, muscle weakness, or impaired oxygenation. When [PO42-] > 2.0 mg/dL replacement can switch to PO ie Neutra-phos or K-phos (1200-1500 mg daily) but be sure to d/c sucralfate or phosphate-binding antacids. Oral phosphate causes diarrhea which limits its usefulness. Oral maintenance is 1200 mg daily, and IV maintenance is 800 mg daily. Patients in renal failure should be dosed more cautiously

Intravenous PO42- Replacement Therapy

Hyperphosphatemia

Usually due to renal insufficiency (decreased secretion) or widespread cell death (tumor lysis, rhabdomyolysis). Sucralfate or Al(OH)3 antacids can be used to reduce serum levels, and calcium acetate tablets (667 mg) can lower serum phosphate while at the same time raising calcium levels. Hemodialysis is a last resort—-

Potassium

Only 2% of total body potassium is extracellular (0.004% in plasma), and this severely limits the value of serum potassium measurements – the relationship between serum and total body potassium is non-linear such that at [K+] < 3.5, small decreases in serum levels reflect huge decreases in total body K+, whereas at [K+] > 5.0, large increases in serum levels reflect small increases in total body K+ [Kidney Int 30: 116, 1986]

Hypokalemia (transcellular shifts vs. total body depletion)

β2-agonists are known to cause hypokalemia via transcellular shift, but at therapeutic doses this is usually ~ 0.5 mEq/l [Ann Emerg Med 21: 1337, 1992]. On the other hand, when combined with diuretics the additive effects can be significant [Am J Med 86: 653, 1989]. Other causes of transcellular shift into the cell are alkalosis, insulin, and hypothermia.

The leading cause of renally lost hypokalemia is diuretic therapy. In the ICU, additional culprits include NG suctioning, magnesium deficiency [Crit Care Clin 7: 225, 1991], and alkalosis (actually due to shift, see above) – to help differentiate these, analyze the urine chloride (> 25 mEq/L for diuretics or Mg deficiency, < 15 mEq/L for NG drainage or alkalosis). The leading cause of extrarenally lost hypokalemia is diarrhea.

Patients usually asymptomatic until 2.5 mEq/L, at which point look for weakness and mental status changes, as well as > 1 mm U waves, loss of T waves, and QT prolongation. Hypokalemia alone does not cause arrhythmias – rather, when combined with another abnormality (such as dig toxicity, magnesium depletion), it can act to enhance them.

Management should first focus on reversing transcellular shifts (ex. alkalosis, hypothermia). After these have been ruled out, proceed to potassium replacement. KCl is available in 1.5 and 2 mEq/L concentrated solution (ampules of 10, 20, 30, 40 mEq), but these are 4000 mOsm/L and must be diluted [Trissel LA Handbook Inj Drug: 886, 1994]. KPO4 is also available and some prefer it in DKA because of the potassium depletion. Always correct calcium as well because hypokalemia protects against the adverse effects of hypocalcemia.

Standard infusion rate is 20 mEq in 100 mL of isotonic saline over 1 hour, although 40 mEq may be needed if levels < 1.5 mEq/L or serious arrhythmias are present (80 mEq have been given safely [Crit Care Clin 7: 143, 1991]). Up to 20 mEq/hr can be given through a central line (and should be because of venous irritation), but above that the dose should be split between two peripheral lines to avoid the possibility of transient cardiac standstill. Remember that because of the potassium curve, serum levels will be slow to change following hypokalemia – still, if [K+] levels appear refractory, check magnesium because this can be a cause of continued potassium loss [Arch Intern Med 145: 1686, 1985]. ALWAYS PUT PATIENTS ON A CARDIAC MONITOR WHEN GIVING INTRAVENOUS POTASSIUM.

Hyperkalemia (transcellular shifts vs. insufficient excretion vs. overload)

The most common cause is artifact, either due to hemolysis (20% of the time [Arch Intern Med 147: 867, 1987]), distal muscle release secondary to the tourniquet [NEJM 322: 1290, 1990], or release during clot formation in the specimen tube. For this reason, a spurious level in an otherwise normal patient should not invoke treatment until a second lab is confirmatory. If real, this is more dangerous than hypokalemia.

β-blockers and digitalis can cause hyperkalemia via transcellular shifts, as can lack of insulin ([K+] > 7 mEq/L can only be caused by digitalis toxicity). Besides drugs, myonecrosis can cause massive release of potassium but until renal function deteriorates, K+ should be adequately cleared. Acidosis is often blamed for hyperkalemia, but respiratory acidosis is not always accompanied by hyperkalemia [Anesth Analg 76: 680, 1993] and there is no clear evidence that organic acidosis (LA, KA) can produce hyperkalemia [Am J Med 71: 456, 1981; Anesth Analg 76: 680, 1993; NEJM 297: 796, 1977] – it may be that the observed hyperkalemia simply accompanies renal failure or RTA.

Renal insufficiency can produce hyperkalemia if GFR < 10 mL/min or UOP < 1L/day [J Intensive Care Med 3: 52, 1988]. Other renal losses occur in hypoadrenergic states and drugs such as ACE-I, K-sparing diuretics, and NSAIDS (and to a lesser extent heparin, TMP-SMX, cyclosporine, and succinylcholine). When the source is not clear (cellular shift vs. insufficient losses), get a urine potassium (> 30 mEq/L suggests shift, < 30 mEq/L suggests insufficient loss). After 14 days, 1 unit of PRBC provides a 3.1 mEq potassium load, and a unit of whole blood provides 4.4 mEq [Transfusion 15: 144, 1975]. This is adequately cleared except in circulatory shock, ie massive resuscitations. Because their VD is reduced, these patients can become hyperkalemic.

Cardiac manifestations of hyperkalemia start at 6.0 mEq/L with T waves, elongated PR interval, decreased P wave amplitude, widened QRS (treat rapidly), sine waves, and finally asystole.

To treat critical cases, stabilize with Ca2+(see table for dose depending on clinical state). ONLY GIVE CaCl2 IF QRS WIDE (CaCl2 10cc IVP over 10 min except kids or non-emegencies, they get gluconate). For those taking digitalis, give 10% gluconate diluted in 100 mL isotonic saline over 20-30 min and in those with digitalis toxicity, give MgSO4 instead.

Acute Management of Hyperkalemia

Condition	Treatment	Comment
EKG changes and circulatory compromise (or just wide QRS)	CaCl (10%) 10 mL IV over 3 min	For anyone with wide QRS
EKG changes or K > 7 w/o circulatory compromise	CaGluc (10%) 10 mL IV over 3 min repeat after 5 min if needed	Response lasts ~ 25 min, do NOT give bicarbonate after calcium
AV block refractory to Ca2+	10 U regular insulin in 500 mL of 20% dextrose infused over 1 hr/ Transvenous pacemaker	Should drop K by 1 mEq/L for 1-2h
Digitalis cardiotoxicity	MgSO4 2 g IV bolus; Digibind	Do not use calcium in digitalis toxicity!!!
After the acute phase has resolved or when the EKG stops changing	Kaeyexalate 30 g in 50 mL of 20% sorbitol PO or 50 g in 200 mL 20% sorbital PR	Oral dosing preferred

Drive in with insulin/dextrose and maybe β-agonists, give volume to dilute. Bicarbonate can be tried if acidotic, but note that most patients will be in renal failure, and in this scenario insulin works much better than bicarbonate [Am J Med 85: 507, 1988] and also bicarbonate binds Ca2+ and should not be given in patients who received it. For these reasons, bicarbonate is rarely useful. Remove with Kayexelate and dialysis. Note that with Kayexalate, every 1 mEq of K lowered will raise Na by 2-3 mEq. Hemodialysis is the most effective method of removal in patients in renal failure [Am J Med 85: 507, 1988]. Lasix, Mg2+, and bicarbonate are rarely used, because these patients are so often in renal failure.

Sources

Miller 5th ed