CHEST PAIN, ACUTE                                HOSPITALIZATION                              MI

DIFFICULT-TO-CONTROL HTN            HYPERTENSION, RESISTANT            MISSED DIAGNOSIS

FIXED-COMBINATION DRUGS            LENGTH OF STAY                                 THROMBOLYSIS
 

THE EVALUATION OF ACUTE CHEST PAIN: This is a difficult subject to review in any article of reasonable length, but NEJM has recently taken a shot at it. ⁽¹⁾ Among the more interesting points are:

1. The ECG is the most important single source of data and should be obtained within 5 minutes after presentation. The prevalence of acute MI (in a series of over 2000 patients) was 80% among patients with 1 mm or more of new ST-segment elevation and 20% among patients with ST-segment depression or T-wave inversion not known to be old. In the absence of ECG changes consistent with the presence of ischemia, the risk of acute MI was 4% among patients with a history of CAD and 2% among patients without such a history.

2. In a typical population of patients with acute chest pain who present at the ED, approximately 15% have acute MI and about 30-35% have unstable angina. 

3. For patients who present within a few hours after onset of symptoms and have ST-segment elevation > 1 mm in 2 or more leads, urgent PTCA or IV thrombolytic therapy is indicated. For patients with other ECG changes that are indicative of ischemia, such as flat or down-sloping ST-segment depression or T-wave inversions, the presumptive diagnosis, until proved otherwise, must be an ongoing acute coronary syndrome with no ST-segment elevation. Treatment with aspirin, IV heparin, or a platelet glycoprotein IIb/IIIa receptor antagonist (and commonly, in addition, nitrates, beta-blockers, or both) should be instituted while the diagnostic evaluation is beginning.

4. Levels of CK-MB usually rise above the normal range within 4 hours after the onset of MI and serial sampling of CK-MB over a period of 12 to 24 hours permits the detection of virtually all acute MIs. The assay for CK-MB is inexpensive and permits the detection of reinfarction. CK-MB levels are not helpful in determining the prognosis in patients with unstable angina.

5. The cardiac troponins rise within about 4 hours and remain elevated for several days. Once elevated, the troponins are not useful in detecting repeated episodes of myocardial ischemia. Multiple studies have shown that elevated levels of troponins indicate an increased risk of complications in patients who do not meet other clinical criteria for acute MI. The interpretation of these results is not always straightforward. Results for troponin that are falsely positive in patients with no evidence of ischemia are generally attributed to nonischemic or subclinical ischemic myocardial injury. False negative results may occur in patients who subsequently have life-threatening complications due to rupture of plaque or arrhythmia. The most appropriate strategy for the combined use of available serum markers remains uncertain. A recent analysis suggested that CK-MB should continue to be the first -line test for patients with suspected ischemic heart disease, and an assay for troponin I or T should be reserved for intermediate-risk patients who have normal CK-MB levels but ECG changes consistent with the presence of ischemia.

6. The assessment of risk:
 

High Risk	Intermediate Risk	Low Risk
prolonged (>20 min) ongoing pain at rest	prolonged (>20 min) angina at rest, now resolved, with moderate or high likelihood of CAD	increased frequency, severity, or duration of angina
pulmonary edema, most likely related to ischemia	angina at rest (> 20 min or relieved with rest or SL NTG)	angina provoked at a lower threshold
angina at rest, with dynamic ST-segment changes > 1 mm	nocturnal angina	new-onset angina with onset 2 wk to 2 mo before presentation
angina with new or worsening mitral regurgitant murmur	new onset Canadian Cardiovascular Society class III or IV angina in the previous 2 week with a moderate or high likelihood of CAD	normal or unchanged ECG
angina with S3 or new or worsening rales	Pathologic Q waves of ST-segment depression < 1 mm in multiple leads at rest.
	age > 65 years

 

AHCPR guidelines state that patients with unstable angina who are at low risk for acute MI do not necessarily require admission.

7. Studies have shown that patients who have a low clinical risk of complications can safely undergo exercise testing within 6 to 12 hours after presentation at the hospital, or even immediately, and that patients with negative tests have excellent outcomes at 6 months. This does not apply, however, to patients with ECG changes consistent with ischemia or those who have infarction not known to be old, ongoing chest pain, or evidence of CHF.

8. A newer approach is to consider prompt coronary angiography in patients who do not meet criteria for acute MI despite suggestive symptoms. Recent analyses indicate that this strategy is particularly cost effective in patients with a high probability of CAD. 
 

HOW OFTEN ARE WE GOING TO MISS THE DIAGNOSIS OF ACUTE MI? These authors studied the incidence of factors related to and the clinical outcomes of failure to hospitalize patients with acute cardiac ischemia. They analyzed all patients with chest pain or other symptoms suggesting acute cardiac ischemia who presented to the ED of 10 U.S. hospitals. Of 10,689 patients, 17% ultimately met the criteria for acute cardiac ischemia (8% had acute MI, 9% had unstable angina), 6% had stable angina, 21% had other cardiac problems, and 55% had noncardiac problems. Among the 889 patients with acute MI, 2.1% (19) were mistakenly discharged from the ED; among the 966 patients with unstable angina, 2.3% (22) were mistakenly discharged. 53% of the patients with a missed diagnosis of acute MI had normal or nondiagnostic ECGs as did 62% of the patients with a missed diagnosis of unstable angina. On follow-up, 74% of missed acute MIs were found to be non-Q-wave MIs that could not be ascribed to a specific location. Multivariable analysis showed that women < 55 years old, nonwhite patients, patients who reported shortness of breath as their chief symptoms, or patients who had normal or nondiagnostic ECGs were more likely not be hospitalized. The authors state, "Efforts to reduce the number of missed diagnoses are warranted." ⁽²⁾

COMMENT: What they don't say is "How?" Although there was disagreement over the interpretation of 16% of the ECGs on subsequent review by an experienced cardiologist, this was not believed to be clinically significant in any of the cases. The authors acknowledge, "Further reduction in the current relatively low rate of missed diagnoses of acute myocardial infarction and unstable angina will be difficult...Whether the availability of a variety of techniques for the diagnosis of acute cardiac ischemia, such as serial measurement of cardiac enzymes, noninvasive cardiac imaging, and predictive instruments, or the use of "chest pain programs" will help reduce the number of missed diagnoses of myocardial infarction or unstable angina is still an unanswered question. It is noteworthy that, in this study, the presence of a well-established chest-pain unit was not related to lower rates of missed diagnosis of acute cardiac ischemia." Fortunately, the risk-adjusted mortality ratio for the nonhospitalized patients with acute MI was not significantly different from those hospitalized. 

My ultimate reaction to this article is to say, "Give me a break!" If ECGs and enzyme are nondiagnostic and the error rate in reading ECGs is not significant, what are we going to do? They say something about eliciting a better clinical history of the chest pain, but they have no data to suggest this will change outcomes. This is not a perfect business. I think we are just going to have to accept an error rate of between 2.1-2.3% unless we want to admit all women and non-Caucasian patients with chest pain. Certainly they don't offer the recipe for better outcomes.
 

THE MANAGEMENT OF DIFFICULT-TO-CONTROL HYPERTENSION: There is a linear relationship between blood pressure and both risk of stroke and rate of decline in renal function due to all forms of renal disease. Optimal blood pressure is less than 120/80 mm Hg in regard to cardiovascular risk. Lowering the blood pressure as far below 140/90 mm Hg as can be tolerated by the patient may provide additional benefit. BP control is often not very effective, even in the trials. A goal diastolic BP < 90 mm Hg was reached in only 72% of subjects in the TOMHS trial. Reduction of systolic BP to a goal < 140 mm Hg is achieved even less often. In the ALLHAT trial only 28% of hypertensives had BP < 140/90 Hg at entry into the study. Two years later, 60% of patients had BP < 140/90 mm Hg. In the recent Systolic Hypertension in Europe trial, only 43.5% of patients reached a systolic BP < 140 mm Hg. Recent treatment trials have shown that one of the myths of hypertension treatment is that most patients achieve control with only a single agent. The HOT and Syst-Eur trials showed that, at most, only one half of patients achieve goal blood pressure with a single agent. Another third of patients require 2 agents, and the rest need 3 or more agents. 

Resistant hypertension is a failure to decrease BP to less than 140/90 mm Hg while the patient is taking 3 or more appropriate antihypertensive agents. In resistant hypertension, attention to sodium intake and basing the antihypertensive regimen on aggressive diuretic therapy (HCTZ, 50 mg/d; metolazone, 2.5-5 mg/d; furosemide, 40-120 mg/d; bumetanide, 1-4 mg/d) often lead to success in reaching goal blood pressure. Other factors to consider are cigarettes and coffee-a recent study of 16 hypertensives showed that the equivalent of smoking 2 cigarettes and drinking 2 cups of coffee increase systolic and diastolic BP a mean of 16 mm Hg and 9 mm Hg, respectively. Smoking 2 packs of cigarettes and/or drinking 6 to 8 cups of coffee elevates blood pressure throughout the waking day. In addition patient use of alcohol and NSAIDs should be carefully noted. Unfortunately, patient education has not been effective in significantly improving the achievement of blood pressure goals in over 20 years of national efforts by the NHLBI. 

This author ⁽³⁾ recommends the following strategies for this group of resistant patients.

1) Use more fixed combination drugs. Diuretics plus beta-blockers, ACE-inhibitors, and angiotensin II blockers have long been available in fixed combination. Now more aggressive combinations are available: calcium channel blockers with ACE-inhibitors, e.g., felodipine and enalapril (Lexxel), amlodipine and benazepril (Lotrel), diltiazem and enalapril (Teczem), and verapamil and trandolopril (Tarka).

2) Schedule follow-up visits at increased frequency. The HOT trial and the Syst-Eur trial suggest that frequent visits with aggressive titration of BP are essential to achieve goal blood pressure. Maintenance of blood pressure control requires frequent patient visits, at least every 3 months. Home monitoring can further result in substantial improvement in blood pressure control. Rapid titration of BP medications may also be effective. COMMENT: One good treatment model to follow is that used in the HOT study, where they achieved an overall goal of 92% reduction of diastolic BP to < 90 mm Hg. Drug treatment consisted of first-line treatment with 5 mg felodipine (Plendil ) q.d. , then either an ACE-inhibitor or a beta-blocker was added as second-line therapy, if needed, and third-line treatment consisted of an increase in felodipine to 10 mg q.d., and fourth-line treatment of doubling the dose of either the ACE-inhibitor or the beta-blocker. Diuretic therapy was reserved until the fifth step. We generally consider the treatment of hypertension to be rather simple. It's not.
 

DON'T KEEP YOUR MI PATIENTS TREATED WITH THROMBOLYTICS IN THE HOSPITAL FOR MORE THAN 3 DAYS: In a study of 22,361 patients with acute MI with an uncomplicated course for 72 hours after thrombolysis, these authors examined the cost effectiveness of an additional day of hospitalization in this group. ⁽⁴⁾ Of the patients with an uncomplicated course for 72 hours after thrombolysis, 16 had ventricular arrhythmias during the next 24 hours; 13 survived for at least 24 hours. On average, another 0.006 year of life per patient could be saved by keeping patients with an uncomplicated course in the hospital another day at a cost of $105, 629 per year of life saved. In a sensitivity analysis, it was found that a fourth day of hospitalization would be economically attractive only if its cost could be reduced by more than 50% or if a high-risk subgroup could be identified in which the estimated survival benefit would be doubled.
 
 
 

Don't forget to visit us on the Web at www.pcpress.net. You'll find all issues, all years, of all newsletters, complete with CME exams.

DEHYDRATION                                           FORMULA FOR SODIUM                         QUALITY OF CARE

ELECTROLYTE DISORDERS                    HYPERNATREMIA                                    SODIUM FORMULA

FECAL IMPACTION                                     HYPONATREMIA                                      SODIUM SOLUTIONS
 

HYPERNATREMIA IS GETTING TO BE MORE OF A PROBLEM: In a prospective study from a 942-bed urban university hospital, the authors sought to determine the incidence, clinical character- istics, and outcome for general medical-surgical patients with hypernatremia (Na > 150 mmol/L) over a 3-month period. ⁽⁵⁾ They identified 103 patients, of whom 18 were hypernatremic on admission and 85 developed hypernatremia during hospitalization. Patients who developed hypernatremia during hospit- alization were younger than patients who developed hypernatremia before hospital admission. Eighty-nine percent of patients who developed hypernatremia during hospitalization had urine concentrating defects, primarily as the result of the use of diuretics or of solute diuresis, whereas only 50% of patients who were hypernatremic on admission could be shown to have concentrating defects. 55% of all hypernatremic patients had increased insensible water losses, and 35% had increased enteral water losses. 86% of patients with hospital acquired hypernatremia lacked free access to water, 74% had enteral water intake of < 1 L/d, and 94% received less than 1 L of IV electrolyte-free water per day during the development of hypernatremia. No supplemental electrolyte-free water was prescribed during the first 24 hours of hypernatremia in 49% of patients. The duration of hypernatremia was shorter in patients who were hypernatremic on admission (median, 3 days) than in patients with hospital-acquired hypernatremia (median, 5 days). Hypernatremia was estimated to contribute to mortality in 16%. The authors conclude that hospital-acquired hypernatremia is primarily iatrogenic resulting from inadequate and inappropriate restriction of fluids to patients with predictably increased water losses and impaired thirst or restricted free water intake or both. They also comment that treatment of hypernatremia is often inadequate or delayed. COMMENT: We are certainly seeing an increase in frequency of this electrolyte disorder on our inpatient service-primarily among the elderly who free water needs have either not been assessed or has been ignored.
 

HYPERNATREMIA IN THE ICU: AN INDICATOR OF QUALITY OF CARE? This was a retrospective survey from a university teaching hospital of all patients (N=389) who were admitted to the medical ICU during a one year period. Of the 389 patients, hypernatremia was present in 34 (8.9%) on admission, and the average duration of hypernatremia was 16.2 hours. A total of 22 patients (5.7%) developed hypernatremia in the course of their stay, and the average duration of hypernatremia in this group was 34.7 hours. Moderately elevated levels of sodium had been detected in most of these patients (n=21) in the days before the development of severe hypernatremia, but adjustments in fluid infusion aimed at preventing the occurrence of hypernatremia were either lacking (n=7), or inadequate (n=11). Hospital-acquired hypernatremia was associated with a higher mortality rate than hypernatremia present on admission to the ICU (32% vs 20.3%). These authors believe that hypernatremia could potentially be used as an indicator of quality of care in the medical ICU. COMMENT: It is quite surprising that we don't do better in our management of this common electrolyte abnormality than we do. I suspect that it is because we're confused about sodium and free water metabolism and because previous guidelines have not been applicable to the full range of common situations. Fortunately, a very simple, concise review of the management of hypernatremia has just appeared in NEJM. ⁽⁶⁾ But first, a relevant case example.
 

DEATH FROM FECAL IMPACTION: An 87 year old mildly demented woman was admitted from a nursing home for acute, heavy bleeding per rectum. This patient had had a previous admission two months previously for acute back pain due to a diagnosis of compression "burst" fracture of L3 after a fall. Her pain had been difficult to control despite calcitonin and low-level narcotics. She was discharged to the nursing home on Oxycontin b.i.d., continued calcitonin, and a clam-shell brace for her spine. Her pain was never adequately controlled. In the NH they had progressive problems with feeding and declining mental status. A dysphagia evaluation had resulted in a recommendation for pureed feedings. Because she had pain with any change of position she was fed in the recumbent position. The nurses reported increasing difficulty with feeding and concentrated on only getting the pureed food down; they stopped offering water and clear liquids afterwards because the patient would tire. The patient would frequently grab her abdomen as if in pain after meals. Her mental status was becoming increasing disoriented and unclear. On the day of admission she had an episode of thick mushy blood per rectum that oozed continually and was referred for admission. In the ED her sodium was 172 mEq/L.

What happened? There was a delay in recognition of critical hypernatremic dehydration despite the presence of several risk factors: an L-S spine film on her prior admission had shown fecal impaction that was not recognized; the Oxycontin and continued immobility increased the risk of progression with the attendant risk of abdominal pain, nausea, and decreased appetite. When this occurred, the nursing home staff paid preferential attention to the intake of a pureed diet at the expense of free water. Finally, whatever her bowel disease was (she died before it was diagnosed) increased hypotonic fluid losses.
 

Hypernatremia is defined as a rise in serum sodium > 145 mmol/L. It represents a deficit of water in relation to the body's sodium stores, which can result from a net water loss or a hypertonic sodium gain, but net water loss accounts for the majority. It can occur in the absence of a sodium deficit (pure water loss), or as a hypotonic fluid loss (as with persistent vomiting or NG suction). Hypernatremia always indicates hypertonic hyperosmolarity and always causes cellular dehydration (at least transiently). Sustained hypernatremia can occur only when thirst or access to water is impaired, the groups at highest risk are patients with altered mental status, intubated patients, infants, and elderly persons. Note particularly that thirst impairment occurs commonly in elderly persons; frail nursing home residents and hospitalized patients are prone to hypernatremia because they depend on others for their water requirements. Note also that there are few symptoms, especially in the elderly. While intense thirst may be present initially, it dissipates as the disorder progresses. Hypernatremia is frequently iatrogenic. 

Management requires a 2-pronged approach: 1) address the underlying cause (e.g. GI fluid losses, fever, hyperglycemia, glucosuria, lactulose therapy, diuretic therapy, hypercalcemia, hypokalemia, or correcting the feeding preparation); and 2) correct the hypertonicity.

If the patient's hypernatremia has developed rapidly (over hours), you can correct it rapidly without a risk of cerebral edema; aim for a reduction of sodium of 1 mmol/hour. For patients with hypernatremia or longer or unknown duration, correction must be carried out more slowly because the full dissipation of accumulated brain solutes occurs over a period of several days. In such patients the sodium should be reduced at a maximal rate of 0.5 mmol/L/hour to prevent cerebral edema and seizures. These authors recommended a target fall in serum sodium of 10 mmol/L/d for all patients with hypernatremia except those in whom the disorder has developed over a period of hours.

Only hypotonic fluids are appropriate; these include pure water (Na=0), 5% dextrose (Na=0), 0.2% sodium chloride (Na=34 mmol/L); 0.45 sodium chloride (Na=77 mmol/L); Ringer's lactate (Na=130 mmol/L), and normal saline (Na=154 mmol/L). The more hypotonic the infusion, the lower the infusion rate required. Because the risk of cerebral edema increases with the volume of the infusate, the volume should be restricted to that required to correct hypertonicity. Normal saline (0.9% sodium chloride) should not be used in the management of hypernatremia unless dehydration is present to an extent that is causing hemodynamic compromise. 

These authors provide a useful formula for determining which hypotonic solution to use:

Change in serum Na=(infusion Na - serum Na) ÷ (total body water + 1)

or Change in serum Na=(infusion Na + infusion K - serum NA) ÷ (total body water + 1) (the + 1 is for each liter of fluid that you administer). The estimated total body water (in liters) is calculated as a fraction of body weight. The fraction is 0.6 in children; 0.6 and 0.5 in nonelderly men and women, respectively, and 0.5 and 0.45 in elderly men and women, respectively. Normally extracelluar and intracellular fluids account for 40 and 60% of total body water, respectively. 

Thus for the patient above, after initial hydration to adequate volume status, the appropriate fluid and rate to have used would have been (using .2% sodium chloride): 0.34 (infusate Na)-172 (patient's Na) / 50 kg (patient's weight) x .45 (TBW as fraction of weight; for elderly women fraction=0.45) + 1 (for each liter of fluid given). Change in serum Na= 34 - 172 ÷ 50 + 1 = - 138/51 = - 2.7 mEq. Thus to reduce the patient's sodium by the maximal 10 mEq/d, or 5 mEq per 12 hours, at least 2 liters of fluid must be given, and another 1 liter should be added for ongoing losses of gastric, intestinal and other fluids, so a total of 3 liters of .2% saline should be given over 12 hours, or approximately 250 mL per hour. If it is desired to give less total volume of fluid or to lower the Na more rapidly, a more dilute infusion solution can be used.

As an example of how normal saline is inappropriate, consider a 50-year old man with Na of 162 mmol/L and a weight of 70 kg. The retention of 1 liter of 0.9% sodium chloride will decrease the serum sodium by only 0.2 mmol/L. Even though the sodium concentration of the infusate is lower than the patients' serum sodium, it is not sufficiently low to alter the hypernatremia substantially. 

They do not recommend using the older traditional formula (water deficit=total body water x [1-(140/Na]) because, while it provides an adequate estimate of the water deficit in patients with hypernatremia caused by pure water loss, it underestimates the deficit in patients with hypotonic fluid loss; nor is it useful when sodium and potassium, in addition to water, must be prescribed. 
 

THE MANAGEMENT OF HYPONATREMIA: The same authors tackle the problem of hyponatremia. Hyponatremia is defined as a decrease in serum sodium to < 136 mmol/L. Hyponatremia can be associated with low, normal, or high tonicity. Examination of the concentration of the patient's urine will help guide therapy. Effective osmolality or tonicity refers to the contribution to osmolality of solutes, such as sodium and glucose, that cannot move freely across cell membranes, thereby inducing transcellular shifts in water. Dilutional hyonatremia is by far the most common form of this disorder in which water intake exceeds the capacity of the kidneys to excrete water, usually by excessive water intake in the presence of normal or near normal kidney function. Conditions with impaired water excreting capacity are most commonly due to ADH. The nonhypotonic hyponatremias are hypertonic (or translocational) hyponatremia, isotonic hyponatremia, and pseudohyponatremia (usually caused by severe hypertriglyceridemia or paraproteinemia). Translocational hyponatremia results from a shift of water from cells to the ECF that is driven by solutes confined in the extracellular compartment (e.g., glucose or mannitol). [For each increase of 100 mg/dL of glucose the sodium will decrease by approximately 1.7 mmol/L.] Most patients with serum sodium > 125 mEq/L will have no symptoms. The most common causes of severe hyponatremia in adults are therapy with thiazides, the postoperative state, and other causes of the syndrome of inappropriate ADH, polydipsia in psychiatric patients, and transurethral prostatectomy (involving extensive irrigant solutions without sodium).

There is no consensus about the optimal treatment of symptomatic hyponatremia. These authors suggest that correction should be of a sufficient pace and magnitude to reverse the manifestations of hypotonicity but not be so rapid and large as to pose a risk of development of osmotic demyelination. A relatively small increase in the serum sodium concentration (e.g., 5%) should substantially reduce cerebral edema. Even seizures induced by hyponatremia can be stopped by rapid increases in the serum sodium that average only 3 to 7 mmol/L. Most reported cases of osmotic demyelination occurred after rates of correction that exceeded 12 mmol/L/day. These considerations suggest a target rate of correction not exceeding 8 mmol/L on any day of treatment.

You can use the same formula as above to calculate the type of fluid and rate to be administered. Change in Na= (infusion Na - serum Na) ÷ (total body water + 1), or with potassium containing fluids, or Change in Na=((infusion Na + infusion K) - serum Na ) ÷ (TBW + 1). Available solutions include: Ringer's Lactate with 130 mmol/L Na, normal saline with 154 mmol/L Na, 3% hypertonic sodium chloride with 513 mmol/L Na, and 5% hypertonic saline with 855 mmol/L Na.

For patients with symptomatic hypotonic hyponatremia with concentrated urine (> 200 mOsm/kg H2O) and clinical euvolemia or hypervolemia, infusion of hypertonic saline is required, usually combined with furosemide to limit treatment induced expansion of the ECF. Furosemide-induced diuresis is equivalent to a one-half isotonic saline solution and thus aids in the correction of hyponatremia. Patients with hypovolemia can simply be treated with isotonic saline.

Patients with symptomatic hyponatremia and dilute urine (osmolality < 2000 mOsm/kg H20) with less serious symptoms usually require only water restriction and close observation. ⁽⁷⁾ COMMENT: These two articles are a very nice, simple summary of an otherwise confusing subject that is frequently mismanaged. I recommend getting copies of both articles for your permanent files and keeping a hawkish eye out for the insidious development of hypernatremia in your elderly patients.

1. Lee TH, Goldman L. Evaluation of the patients with acute chest pain. N Engl J Med 2000; 342: 1187-95.

2. Pope JH, Aufderheide TP, Ruthazer R, Woolard RH, Feldman JA, Beshansky JR et al. Missed diagnoses of acute cardiac ischemia in the emergency department. N Engl J Med 2000; 342: 1163-70.

3. Graves JW. Management of difficult to control hypertension. Mayo Clin Proc 2000; 75: 278-84. THE HOT STUDY: Hansson L, Zanchetti A, Carruthers SG, Dahlof B, Elmfeldt D, Julius S et al. Effects of intensive blood-pressure lowering and low-dose aspirin in patients with hypertension: principal results of the Hypertension Optimal Treatment (HOT) randomised trial. Lancet 1998; 351: 1755-62

4. Newby LK, Eisenstein EL, Califf RM, Thompson TD, Nelson CL, Peterson ED et al. Cost effectiveness of early discharge after uncomplicated acute myocardial infarction. N Engl J Med 2000; 342: 749-55.

5. Palevsky PM, Bhagrath R, Greenberg A. Hypernatremia in hospitalized patients. Ann Intern Med 1996; 124: 197-203.

6. Adrogue HJ, Madias NE. Hypernatremia. N Engl J Med 2000; 342: 1493-99.

7. Adrogue HJ, Madias NE. Hyponatremia. N Engl J Med 2000; 342: 1581-89.