Electrolyte disorders are commonly encountered in the emergency department. Studies investigating the prevalence and relevance of disorders in serum sodium and potassium exist both for hospitalized noncritically ill patients and for critically ill patients 1–3. Interestingly, although studies showed that both disorders of serum sodium as well as potassium are independently associated with adverse outcome in various patient populations, only little data have been published on dysnatremias and dyskalemias in the emergency room setting 2,4–7. So far, one study from our group has investigated the prevalence and clinical significance of severe hyponatremia and hypernatremia in patients presenting to the emergency department 8. A study from Taiwan only focused on the prevalence of hyponatremia in the emergency department 9. Another investigation from our group aimed at identifying the effect of thyroid hormones on electrolyte levels in patients admitted to the emergency room 10. However, the study did not evaluate the origin and clinical significance of hypokalemia in this patient population. Thus, hypokalemia never seems to have been investigated in a large patient collective admitted to the emergency room – only some case reports exist.
Given the potential high morbidity associated with hypokalemia, ranging from muscle weakness, rhabdomyolysis, and functional and structural abnormalities of the kidney to cardiac arrhythmias and ECG changes, a systematic study investigating the prevalence and clinical significance of hypokalemia in the emergency setting is desirable 11–13.
Thus, we aimed to investigate (a) the prevalence of hypokalemia in the emergency department, (b) the etiology of severe hypokalemia, and (c) the symptoms of severe hypokalemia including ECG changes.
Study design and setting
In this retrospective analysis, all patients admitted to the Emergency Department of the Inselspital, University Hospital Bern, between 1 April 2008 and 31 March 2011 (3 years) with measurement of serum potassium were included.
Selection of participants and methods
Using these data, we calculated the prevalence rates for hypokalemia and hyperkalemia defined according to the reference range of our central laboratory. Hypokalemia was defined as a serum potassium level below 3.5 mmol/l and hyperkalemia as a serum potassium level exceeding 4.5 mmol/l. Hyponatremia was defined as a serum sodium level below 135 mmol/l and hypernatremia as a serum sodium exceeding 145 mmol/l.
Of all patients with severe hypokalemia, defined as a serum potassium level of 2.5 mmol/l or below, patient charts from the emergency department as well as from the subsequent hospitalization were gathered. A thorough chart review was performed by two physicians (G.L. and G.M.) to obtain information on the symptoms of patients presenting with severe hypokalemia, the changes observed in the ECG performed in the emergency department, the initial treatment for hypokalemia in the emergency department as well as the etiology of hypokalemia. Moreover, hospitalization charts were reviewed to obtain information on correction of hypokalemia during hospitalization and the final evaluation of the etiology of hypokalemia.
In addition, we gathered laboratory data on serum electrolytes, creatinine, and venous blood gas analysis of all patients presenting with severe hypokalemia.
Only symptoms not explainable by a pathology other than hypokalemia were attributed to the electrolyte disorder. All case charts were reviewed by two physicians experienced in the diagnosis and treatment of electrolyte disorder. In case of discordance in terms of symptoms and etiology, a third senior physician was consulted.
Only the first admission during the study period was considered for the analysis in patients who presented more than once with severe hypokalemia.
The database was created using Excel (Microsoft Corp., Redmond, Washington, USA). Statistical analysis was carried out using IBM SPSS Statistics 19.0 (International Business Machines, Armonk, Virginia, USA).
All data are presented as means with SD or medians and first and third quartiles as appropriate. χ2-Test, Fishers exact test, t-test, and Mann–Whitney U-test were used to analyze for differences between groups.
The study was approved by the local institutional review board (Ethic commission of the Canton of Bern, Switzerland, http://www.kek-bern.ch).
During the study period, serum potassium measurements for a total of 43 805 patients were ordered by the Department of Emergency Medicine. The median age of the patients on admission was 53 years (Q1: 36, Q3: 70). The median serum creatinine was 72 μmol/l (Q1: 60, Q3: 88), potassium was 3.9 mmol/l (Q1: 3.7, Q3: 4.2), and sodium was 139 mmol/l (Q1: 137, Q3: 141).
4826 patients (11%) had a serum potassium below 3.5 mmol/l, defining hypokalemia. 3903 patients (9%) presented with hyperkalemia defined as a serum potassium exceeding 4.5 mmol/l. An overview of electrolyte disorders stratified for serum potassium is presented in Table 1. The distribution of serum potassium is presented in Fig. 1.
A total of 53 patients with severe hypokalemia (serum potassium <2.6 mmol/l) could be identified during the study period. Thirty-four patients (64%) were triaged as medical and 19 (36%) as surgical patients. Neurologic causes for emergency department admission were present in 17 patients (32%), gastrointestinal in 11 (21%), trauma in seven (13%), an unclear deterioration of general health state in seven (13%), infection and hypokalemia in three (5%), respectively, respiratory in two (5%) and intoxication, malignancy, and cardiac in one patient (2%), respectively. The median age of the patients was 56 years (Q1: 37, Q3: 73), 36 women (68%) and 17 men (32%). Women were significantly over-represented in the group with severe hypokalemia compared with the normokalemic group (68 vs. 42%, P=0.0002). There was no difference in terms of age between patients with severe hypokalemia and patients with normokalemia (53 vs. 56 years, P>0.05). The median serum creatinine was similar in patients with normokalemia and those with severe hypokalemia (72 vs. 69 μmol/l, P>0.05) (Fig. 2).
In patients with severe hypokalemia, the median serum potassium was 2.4 mmol/l (Q1: 2.2, Q3: 2.5), with a minimum of 1.5 mmol/l. The median serum creatinine was 69 μmol/l (Q1: 44, Q3: 110) and the median serum sodium was 139 mmol/l (Q1: 134, Q3: 142). Venous blood gas analysis was carried out in 17 patients (32%) with severe hypokalemia. The median pH was 7.46 (Q1: 7.32, Q3: 7.5), pCO2 35 mmHg (Q1: 30, Q3: 38), and bicarbonate 22 mmol/l (Q1: 18.6, Q3: 24.6). Five patients showed a picture of respiratory alkalosis, four of metabolic alkalosis, three of respiratory acidosis, two of metabolic acidosis, and one of combined respiratory and metabolic alkalosis in the venous blood gas analysis. Only two patients had a completely normal venous blood gas analysis.
Hyponatremia was significantly more common in patients with severe hypokalemia than in patients with normal serum potassium (30 vs. 10%, P<0.0001) or milder forms of hypokalemia (i.e. serum potassium between 2.5 and 3.4 mmol/l; 30 vs. 13%, P=0.0009). Also, patients with severe hypokalemia were more likely to have hypernatremia than patients without hypokalemia (8 vs. 2%, P=0.003), but there was no significant difference compared with patients with milder forms of hypokalemia (8 vs. 3%, P=0.07).
Symptoms not explainable by a pathology other than hypokalemia were present in 26 patients (49%). Although five patients presented with two symptoms attributed to severe hypokalemia, 21 patients had one symptom. Weakness, present in 19 patients (73% of symptomatic and 36% of all patients), was the leading symptom of hypokalemia. Muscle pain was present in four patients (15/8%), obstipation in three (12/6%), muscle cramps in two (13/4%), and cardiac arrhythmia (ventricular fibrillation), collapse, and seizures in one patient (4/2%), respectively.
No symptoms because of hypokalemia were found in 15 patients (28%) and in 12 patients (23%), no statement in terms of hypokalemia-related symptoms could be made because of a reduced state of consciousness of these patients and consequent absence of a detailed patient history.
In 29 patients (55%), the ECG that was performed during the emergency department stay was available. Of these, 20 patients (69% of patients with ECGs performed in the emergency department) showed ECG changes because of hypokalemia. Twelve patients had one single ECG change, six patients had two changes, and two patients had three different changes in the ECG performed on admission to the emergency department. The presence of a U wave was the most common ECG sign of hypokalemia as present in seven patients (24%), followed by ST segment depression and the presence of multiple ventricular extrasystoles in six patients (21%), respectively. ECG changes were not significantly more common in symptomatic than in asymptomatic patients (46 vs. 33%, P=0.519).
Twenty-four patients (45%) received no treatment for hypokalemia during their emergency department stay. One of these patients did not receive correction as a discontinuation of treatment was decided because of the severity of the underlying disease. All except one patient who received no correction during their emergency department stay were hospitalized. Twenty-three patients (43%) received intravenous potassium replacement; four (8%) of these received an additional substitution of magnesium intravenously and five (9%) an additional oral potassium substitution. Six patients (12%) received oral potassium substitution exclusively.
In 22 patients (42%), no definite etiology of severe hypokalemia could be identified by use of the emergency department chart. In five of the patients with an unclear etiology of hypokalemia after emergency department chart review, a cause for hypokalemia could be identified using the hospitalization chart following the emergency admission, leaving 17 (32%) with an unknown cause of severe hypokalemia.
Twenty-two patients (42%) had one single cause of hypokalemia, whereas eight (15%) had two causes and five (9%) had three different causes of severe hypokalemia. Malnutrition and use of diuretics were the main causes of severe hypokalemia as present in 15 patients (28%), respectively. Of the 15 patients taking diuretic medication, nine were taking a thiazide-type diuretic and six were taking a loop diuretic. Prolonged vomiting was the third leading cause of severe hypokalemia present in 12 patients (23%), followed by diarrhea, which was present in seven (13%). Laxative abuse was the cause of hypokalemia in two patients (4%) and metabolic alkalosis and a neobladder in one patient, respectively.
Forty-six patients (87%) were hospitalized after referral to the emergency department. The median hospitalization time in patients with severe hypokalemia was 9 days (Q1: 5, Q3: 13). There was no significant difference in the hospitalization time between patients who were symptomatic and those who were asymptomatic [9 (Q1: 5, Q3: 13) vs. 9 (Q1: 5, Q3: 12), P=0.257].
Nine patients (17% of all and 20% of hospitalized patients) died during hospitalization. Causes of death were intracerebral hemorrhage in two patients, decompensated cirrhosis of the liver with development of hepatorenal syndrome in two, rapid progression of an underlying malignancy in two and septic shock, severe head trauma, and malnutrition because of anorexia nervosa in one patient, respectively. There was no significant difference in serum potassium level between patients who died and those who survived [2.4 (Q1: 2.2, Q3: 2.5) vs. 2.4 mmol/l (Q1: 2.2, Q3: 2.5), P=0.349]. No difference could be found in terms of symptoms because of hypokalemia between patients who died and those who survived (44 vs. 49%, P=1.0).
In the present study, including more than 43 000 serum potassium measurements, we found a prevalence for hypokalemia of 11%. Fifty-three patients were identified with severe hypokalemia defined as a serum potassium below 2.6 mmol/l. A significant proportion of patients were identified to have a concurrent disorder of serum sodium balance along with hypokalemia. Surprisingly, only half of the patients with severe hypokalemia showed symptoms attributable to the electrolyte disorder and 69% of patients showed ECG changes because of hypokalemia in the emergency department. The presence of U waves, ST segment depression, and multiple ventricular extrasystoles were the most common ECG changes present in over 20% of patients, respectively. Malnutrition, diuretic medication, and prolonged vomiting were the main causes for severe hypokalemia in our patient collective.
To our knowledge, this is the first study that has systematically evaluated the clinical symptoms and ECG changes in a collective of emergency patients with severe hypokalemia. A recent study from an emergency department in the USA investigated the prevalence of hypokalemia in patients with diabetic ketoacidosis 14. The authors found a prevalence of 6% of hypokalemia in patients with diabetic ketoacidosis. However, because of the special patient population of this study, the prevalence rate cannot be compared with our finding of an 11% prevalence rate for hypokalemia. Most other publications in emergency department patients are case reports 15.
Symptoms because of hypokalemia often do not occur with serum potassium concentrations above 2.5 mmol/l if hypokalemia developed slowly 11. However, it is surprising that only half of our patients with serum potassium levels of 2.5 mmol/l or lower presented with symptoms because of the electrolyte disorder. Muscle weakness, muscle pain, and cramps were the leading symptoms of severe hypokalemia in our study. Muscle weakness is described to occur usually only in patients with a serum potassium below 2.5 mmol/l 16. Muscle cramps and even rhabdomyolysis were described to occur in patients with severe hypokalemia; however, these reports are case reports and no original study exists on this topic 17,18.
Sixty-nine percent of patients with severe hypokalemia showed ECG changes because of the electrolyte disorder. A low serum potassium concentration leads to a more negative resting membrane potential and a decrease in membrane excitability. An increase in the action potential duration and a delay in repolarization result. Typical ECG manifestations of hypokalemia include flattened T waves, prominent U waves, ST segment depression, and a fusion of T and U waves 19. Also, an increase in QRS duration, atrioventricular block, or cardiac arrest may occur. A hint for the relevance of the arrhythmogenic potential of hypokalemia is given by studies that studied the prevalence of hypokalemia in patients who were admitted to a hospital after cardiac arrest 20,21.
Malnutrition, diuretic medication, and prolonged vomiting were the most common causes of severe hypokalemia in our patients. It must be emphasized that a long fasting period is necessary to finally result in hypokalemia, because of the potential of the kidney to reduce potassium output to a minimum of 5–25 mmol 11. However, fasting may be a significant contributor to the development of hypokalemia in patients under diuretic medication, with a reduced capacity of renal counteraction. In patients with prolonged vomiting, the potassium loss leading to hypokalemia does not primarily occur through the gastric secretions but through the kidney: loss of acid because of vomiting leads to metabolic alkalosis. Sodium bicarbonate is transported to the distal potassium secretory site. In addition, aldosterone secretion is induced by vomiting. These two effects increase renal potassium excretion and favor the development of hypokalemia 22.
Diuretics are a mainstay of therapy for a multitude of diseases. Thiazide-type diuretics are often part of an antihypertensive therapy regimen and hypokalemia was described to occur in as many as 13% of patients in a recent study 6,23. In addition, hypokalemia is a well-known and common side-effect of loop diuretics 11.
So far, it is unclear whether the development of hypokalemia per se has an independent effect on patient outcome. However, recent evidence from patients with myocardial infarction showed that hypokalemia was an independent risk factor for adverse outcome 7. This can be well imagined, given the multitude of effects of hypokalemia on body functions, for example its proarrhythmogenic potential as discussed above. Thus, more studies are required to find optimal prevention and treatment strategies for hypokalemia.
Our study is limited by its retrospective design. It is possible that prevalence rates for hypokalemia are biased by the fact that not all patients received measurement of serum potassium at their emergency visit. However, prevalence rates were strictly calculated for patients with serum potassium measurement only. Also, there is nonstandardized documentation – all signs and symptoms of hypokalemia were likely not documented for each patient. Only 55% of patients underwent an ECG in the emergency department. Thus, it cannot be excluded that the percent of patients with ECG changes is biased.
Hypokalemia is a common problem in the emergency department, with 1 in 10 patients presenting with the electrolyte disorder. Only a minority of patients presenting to the emergency department had severe hyponatremia, with half of them being clinically symptomatic and two-thirds having ECG changes. Future studies are needed on how to optimally treat patients with hypokalemia in the emergency department.
The authors thank Monika Reusser and Heinz von Allmen from the Center of Laboratory Medicine for their help in acquiring the necessary laboratory data for our study.
Conflicts of interest
There are no conflicts of interest.
1. Palevsky PM, Bhagrath R, Greenberg A.Hypernatremia in hospitalized patients.Ann Intern Med1996;124:197–203.
2. Funk GC, Lindner G, Druml W, Metnitz B, Schwarz C, Bauer P, Metnitz PG.Incidence and prognosis of dysnatremias present on ICU admission.Intensive Care Med2010;36:304–311.
3. Darmon M, Timsit JF, Francais A, Nguile-Makao M, Adrie C, Cohen Y, et al..Association between hypernatraemia acquired in the ICU and mortality: a cohort study.Nephrol Dial Transplant2010;25:2510–2515.
4. Polderman KH, Schreuder WO, Strack van Schijndel RJ, Thijs LG.Hypernatremia in the intensive care unit: an indicator of quality of care?Crit Care Med1999;27:1105–1108.
5. Lindner G, Funk GC, Schwarz C, Kneidinger N, Kaider A, Schneeweiss B, et al..Hypernatremia in the critically ill is an independent risk factor for mortality.Am J Kidney Dis2007;50:952–957.
6. Alderman MH, Piller LB, Ford CE, Probstfield JL, Oparil S, Cushman WC, et al..Clinical significance of incident hypokalemia and hyperkalemia in treated hypertensive patients in the antihypertensive and lipid-lowering treatment to prevent heart attack trial.Hypertension2012;59:926–933.
7. Goyal A, Spertus JA, Gosch K, Venkitachalam L, Jones PG, Van den Berghe G, Kosiborod M.Serum potassium levels and mortality in acute myocardial infarction.JAMA2012;307:157–164.
8. Arampatzis S, Frauchiger B, Fiedler GM, Leichtle AB, Buhl D, Schwarz C, et al..Characteristics, symptoms
, and outcome of severe dysnatremias present on hospital admission.Am J Med2012;125:1125.e1–1125.e7.
9. Lee CT, Guo HR, Chen JB.Hyponatremia in the emergency department.Am J Emerg Med2000;18:264–268.
10. Schwarz C, Leichtle A, Arampatzis S, Fiedler GM, Zimmermann H, Exadaktylos AK, Lindner G.Thyroid function and serum electrolytes: does an association really exist?Swiss Med Wkly2012;142:w13669.
11. Rose BD.Clinical physiology of acid–base and electrolyte disorders2001:5 ed..New York:Mcgraw-Hill.
12. Seldin DW, Giebisch GMujais SK, Katz AL.Potassium deficiency.The kidney: physiology and pathophysiology2000.New York:Lippincott Williams & Wilkins;1615.
13. Siegel D, Hulley SB, Black DM, Cheitlin MD, Sebastian A, Seeley DG, et al..Diuretics, serum and intracellular electrolyte levels, and ventricular arrhythmias in hypertensive men.JAMA1992;267:1083–1089.
14. Arora S, Cheng D, Wyler B, Menchine M.Prevalence of hypokalemia in ED patients with diabetic ketoacidosis.Am J Emerg Med2012;30:481–484.
15. Buyukcam F, Calik M, Erkuran MK, Akay H.Hypokalemia and muscle paralysis after low-dose methylprednisolone.Am J Emerg Med2011;29:573.e1573.e2.
16. Mount DB.In: Sterns RH, Emmett M, eds. Clinical manifestations and treatment of hypokalemia. UpToDate 2012.
17. Comi G, Testa D, Cornelio F, Comola M, Canal N.Potassium depletion myopathy: a clinical and morphological study of six cases.Muscle Nerve1985;8:17–21.
18. Shintani S, Shiigai T, Tsukagoshi H.Marked hypokalemic rhabdomyolysis with myoglobinuria due to diuretic treatment.Eur Neurol1991;31:396–398.
19. El-Sherif N, Turitto G.Electrolyte disorders and arrhythmogenesis.Cardiol J2011;18:233–245.
20. Thompson RG, Cobb LA.Hypokalemia after resuscitation from out-of-hospital ventricular fibrillation.JAMA1982;248:2860–2863.
21. Ornato JP, Gonzalez ER, Starke H, Morkunas A, Coyne MR, Beck CL.Incidence and causes of hypokalemia associated with cardiac resuscitation.Am J Emerg Med1985;3:503–506.
22. Cheungpasitporn W, Suksaranjit P, Chanprasert S.Pathophysiology of vomiting-induced hypokalemia and diagnostic approach.Am J Emerg Med2012;30:384.
23. Kotchen TA.Antihypertensive therapy-associated hypokalemia and hyperkalemia: clinical implications.Hypertension2012;59:906–907.