Incidence and causes of hypokalemia associated with cardiac resuscitation

Incidence and Causes of Hypokalemia Associated with Cardiac Resuscitation JOSEPH P. ORNATO, MD, EDGAR R. GONZALEZ, PharmD, HELEN STARKE, MD, ANTHONY MORKUNAS, MARTHA R. COYNE, PharmD, CAROL L. BECK, PharmD To further investigate the incidence and etiology of hypokalemia during cardiac arrest, the authors compared data on 74 adult nontrauma cardiac arrest patients (44 men, 30 women, age 66 + 13 years) who had a serum potassium (K+) level documented during or immediately following resuscitation with data on 63 adult controls wlth life-threatening medical emergencies presenting to the emergency department who did not experience arrest. Hypokalemia (serum K+ < 3.6 mEq/l) occurred in 25 arrest patients (34%) compared with nine controls (17%). Serum K+ was not significantly different shortly before (3.7 f 0.4) versus immediately after arrest (3.6 f 0.8) in a small subgroup of patients, making intracellular shifting of K+ because of metabolic events during resuscitation an unlikely etiology. Hypokalemia was associated with a 2.5.fold increase in relative risk for cardiac arrest. Patients who were receiving diuretics without K+ supplementation had the highest risk of arrest (4.4.fold increase). Supplementation of K+ appeared to be protective in patients on diuretics. The authors confirm the association between hypokalemia and cardiac arrest and suggest that this metabolic abnormality may be an important risk factor for cardiac arrest. (Am J Emerg Med 1985;3:503-506)

Hypokalemia has long been incriminated as a trigger for ventricular dysrhythmias in both experimental models and patients with underlying heart disease.re4 Thompson and Cobb5 recently reported that hypokalemia is present in half the survivors of out-of-hospital cardiac arrest caused by ventricular fibrillation (vflb). Diuretic administration did not appear to account for the hypokalemia. The authors concluded that hypokalemia following out-of-hospital arrest was caused by an intracellular potassium shift induced by metabolic events during resuscitation. If this explanation is correct, then hypokalemia is a result of resuscitation rather than a contributing cause of the arrest. Unfortunately, Thompson’s study lacked

From the University braska.

of Nebraska

Medical

Center,

Manuscript received August 13, 1984; revision ruary 5, 1985; revision accepted April 10, 1985.

Omaha,

received

Ne-

Feb-

Address reprint requests to Dr. Ornato at his present address: Medical College of Virginia, MCV Station Box 525, Richmond, VA 23298. Key

Words:

Cardiac

resuscitation,

hypokalemia,

potassium.

information about total body potassium deficit or whether patients on diuretics were also on potassium supplementation. This study was undertaken to 1) compare the frequency of hypokalemia in patients who had a cardiac arrest and had a potassium sample drawn with the frequency of hypokalemia in a control group of patients, and 2) further investigate the etiology of hypokalemia in association with cardiac arrest. METHODS Data were collected prospectively on all cardiac arrest victims consecutively treated at the University of Nebraska Medical Center from December 1982 through January 1984. The study population included only adult patients (17 years of age or older) with cardiac arrest that was not caused by external physical trauma who had a serum potassium levels determined and recorded. The decision to draw a sample for serum potassium was left entirely to the judgement of the senior physician in charge of the resuscitation. Pre-hospital arrests included all cardiac arrests that occurred out-of-hospital with transport of the patient to our emergency department (ED) and arrests that began in our ED. In-hospital arrests included all cardiac arrests that occurred anywhere in the hospital other than the ED. All records of pre-hospital arrests were produced by ED personnel using a highly accurate automated electronic flowsheet system.6 We reviewed both the hospital records and rescue squad report forms to obtain data on all pre-hospital cases. In-hospital records were produced with a manual flowsheet. Physicians trained and certified in advanced cardiac life support directed all in- and out-of-hospital arrests. Radiotelemetry was used to communicate with paramedics in the field. American Heart Association guidelines and protocols for resuscitation were followed. Approximately 3 to 5 minutes were allowed between drug administration if no immediate change in rhythm occurred. The following information was collected on each patient: age, sex, premonitory symptoms of arrest, past medical history, initial arterial blood gases obtained within the first 15 minutes of in-hospital arrest or 503

AMERICAN

TABLE 1.

JOURNAL

Comparison

OF EMERGENCY

MEDICINE

of Cardiac-arrest

and Control

Number Age (mean * SD) Sex (male/female) Medical problems (no.) Coronary heart disease Hypertension Congestive heart failure Diabetes Medication (no.) Diuretic Digitalis Potassium supplement Antiarrhythmic Beta blocker Calcium blocker No prior therapy

n Volume 3, Number 6 n November

Patients

Arrest

Control

P-value

74 652 13 44130

53 60 IT 18 32121

NS NS

33 30 28 18

20 20 14 8

NS NS NS NS

41 24 25 22 11 5 5

28 14 4 4 7 3 14

NS NS < 0.0001 < 0.0003 NS NS < 0.03

-

within the first 15 minutes of ED arrival for pre-hospital arrest, the sequence of rhythms and therapies during resuscitation, and outcome. All arrest data entered by nurses and paramedics and all rhythm strips recorded during the arrest were reviewed by an attending physician at the completion of each arrest and verified for accuracy. For in-hospital arrest victims, we recorded pre-arrest serum potassium concentrations if they had been obtained no more than two hours before the arrest. Successfully resuscitated patients who were hypokalemic received intravenous potassium supplementation. We recorded the amount of potassium and the time required to restore the serum potassium level to 3.6 mEq/l (the lower limit of normal in our hospital laboratory). A control group of all patients with life-threatening medical emergencies who consecutively presented to our emergency department, had a samples for serum potassium drawn, and were admitted during the study period was selected for comparison. The following data was analyzed on these patients: age, sex, medical problem and past medical history, current drug therapy, and serum potassium level. Chi-square test using Yate’s correction, student’s ttest, and calculation of relative risk’ were used for statistical analysis. RESULTS During the study period there were 130 adult cardiac-arrest patients, 74 of whom had a serum potassium levels documented. This group of 74 patients constituted the arrest study population. Of these cardiac arrests, 30 (40%) occurred in the pre-hospital setting and 44 (60%) were in-hospital. The control group was composed of 53 patients. Comparative variables of arrest and control patients are shown in Table 1. 504

1986

Patients with cardiac arrest had a significantly higher rate of potassium supplementation, antiarrhythmic therapy, and no preceding drug therapy than controls. Serum potassium of the cardiac-arrest patients averaged 4.2 & 1.3 mEq/l. Control patients had a lower serum potassium levels (3.9 + 0.6 mEq/l), but the difference was not statistically significant. Hypokalemia (serum potassium level below 3.6 mEq/l) was present in 25/74 (34%) of the arrest patients and 9/53 (17%) of controls (P < 0.06). Hypokalemia was associated with a 2.5-fold increase in relative risk of cardiac arrest (Table 2). Patients who were taking diuretics without potassium supplementation had the highest risk of arrest (4.4-fold increase in relative risk). Potassium supplementation appeared to be protective in patients who were on diuretics. The dysrhythmia seen at the onset of arrest was documented for 71 of the 74 arrest patients who had a serum potassium charted. Patients who were hypokalemic had a significantly higher incidence (P < 0.005) of vfib (21124, or 88%) than patients without hypokalemia (24/47, or 51%). There was no statistically significant difference between those who were hypokalemic and those who were not with respect to the incidence of asystole (l/24 versus 10/47), idioventricular rhythm (l/24 versus 6/47), or supraventricular rhythm (l/24 versus 7/47) as the initial dysrhythmia in the arrest. The amount of time necessary to restore the serum potassium level to the lower limit of normal (3.6 mEq/ 1) in patients receiving diuretics was shorter (P < 0.05) in those receiving than in those not receiving chronic potassium supplementation. More potassium (P < 0.05) was required to restore the serum potassium level to 3.6 mEq/l in patients taking diuretics who were not receiving chronic supplementation than in those who received chronic supplementation. Six in-hospital arrest patients had samples for serum potassium drawn before and after resuscitation. The difference in serum potassium before (3.7 2 0.4 mEq/ 1) and after (3.6 f 0.8 mEq/l) resuscitation was not statistically significant. DISCUSSION Approximately 98% of our total body store of potassium is intracellular, and the remainder is found primarily in the serum.” Reduction in total body potassium is usually caused by decreased intracellular potassium content or loss of lean body mass. Chronic diuretic therapy has been shown persistently to reduce total body potassium by 10-15%.8-‘o The serum potassium concentration partly reflects total body potassium and accounts for most of the clinically important effects of this electrolyte.8-1i Signifi-

ORNATO ET AL W HYPOKALEMIA

cant reduction in the serum potassium level, even without total body depletion, can lead to serious physiological derangement.9*10 The trans-membrane action potential of cardiac pacemaker cells is influenced by the relationship between intra- and extracellular potassium.” Hypokalemia may affect the excitability and metabolic function of many different cells.8fr0 It can trigger a wide spectrum of atrial, junctional, and ventricular dysrhythmias by enhancing automaticity and altering conduction velocity. ‘r~* Hypokalemia alone may not be sufficient to produce a clinically significant dysrhythmia unless underlying myocardial disease is present.’ Studies t3,r4 have failed to link hypokalemia with ventricular dysrhythmias in patients who have uncomplicated hypertension and healthy hearts. In contrast, patients who have ischemic, hypertrophied, or dilated hearts are vulnerable to ventricular dysrhythmias.r5*16 Thompson and Cobb’s5 finding of hypokalemia in half of all pre-hospital vfib arrest victims is interesting, but of minimal importance if it is only a consequence of resuscitation rather than a trigger for the arrest. Our findings confirm the frequent presence of hypokalemia in arrest patients, especially in those whose initial rhythm is vtib. However, we disagree with Thompson and Cobb5 as to whether hypokalemia or cardiac arrest comes first. The higher incidence of vfib as the initial dysrhythmia in our patients with hypokalemia as compared with patients without hypokalemia suggests that hypokalemia might trigger vfib in some patients. We were unable to document a significant drop in serum potassium level in the in-hospital arrest patients who had both pre- and post-arrest serum potassium samples drawn. This finding casts doubt on the suggestion that intracellular shifting of potassium during arrest is responsible for the hypokalemia seen afterwards. Our interpretation that hypokalemia precedes and may trigger cardiac arrest is consistent with recent studies by Nordrehaug and Van der Lippe* and Thomas,17 which document hypokalemia as a precipitant of vfib in patients with acute myocardial infarction. Over half (52%) of our patients who were hypokalemic were taking diuretics. Most (77%) of those who were hypokalemic and taking .diuretics were not taking a potassium supplement. Based on the amount of potassium required to raise the serum potassium of our survivors, the total body deficit of potassium is likely to be modest. This is consistent with the known total body potassium deficit caused by diuretics. 10,18*19 Potassium supplementation appears to be protective. Significantly less time was required to restore the serum potassium to the lower limit of normal in patients on diuretics who were on chronic supplementation (3 -+ 10 hours). For Patients on diuretics not receiving

TABLE2.

Relative Risk

of Hypokalemia

IN RESUSCITATION

for Arrest

Patients

Versus Controls Hypokalemict Diuretic/Potassium Supplement’ No/No Yes/Yes Yes/No Total

(No.)

(No.)

Relative Risk

12133 3/23 10118 25174

4/31 l/4 4118 9153

3.9 0.5 4.4 2.5

Arrest

Controls

* Refers to the chronic use of oral diuretics and potassium plementation. T Hypokalemic = serum potassium < 3.6 mEq/l.

sup-

chronic potassium supplementation, this took longer (10 at 14 hours). More rapid return of the serum potassium to normal in a patient resuscitated from ventricular fibrillation may decrease the risk of a recurrent episode. Although diuretics without potassium supplementation may account for more than half of all cases of hypokalemia associated with resuscitation, the cause of hypokalemia in the remaining cases is unclear. Our information on patient medications in pre-hospital arrest patients was obtained from family or friends. It is possible that some patients were in fact receiving chronic diuretic therapy and that this was not accurately conveyed to our physicians. The actual incidence of hypokalemia associated with cardiac arrest may be higher than we have thus far determined. Acidosis frequently accompanies cardiac arrest and is known to shift potassium out of cells, thus tending to mask pre-existing hypokalemia.20 The initial pH of the arrest patients was 7.36, with 31% who had arterial blood gases recorded having a pH of 7.29 or less. Patients who experienced cardiac arrest had a significantly higher incidence of antiarrhythmic drug usage than controls. The relative risk of arrest was increased 6.6-fold in patients receiving an antiarrhythmic drug. This finding suggests either that our patient population had a higher underlying incidence of cardiac rhythm disturbances before arrest that warranted chronic drug therapy, or that such therapy itself may have increased the risk of arrest. Recent studies*l,** have shown that antiarrhythmic drugs may increase the risk of cardiac arrest in up to 20% of patients on such therapy. We cannot exclude the possibility that this “chronic antiarrhythmic” risk factor may act synergistically with hypokalemia in triggering a cardiac arrest. Based on the high incidence of hypokalemia during cardiac arrest, should intravenous potassium be administered during cardiac resuscitation? We believe that documented hypokalemia in the patient with ventricular fibrillation or tachycardia should be treated 505

AMERICAN

JOURNAL

OF EMERGENCY

MEDICINE

n Volume 3, Number 6 n November

because hypokalemia can initiate or perpetuate ventricular dysrhythmias in patients with underlying ischemit heart disease.2~‘*,12,‘5-‘7However , we do not advacate giving potassium empirically without knowledge of the serum level (as might be the case in an out-of-hospital arrest run by paramedics), because hyperkalemia cannot be ruled out. Hyperkalemia (serum potassium above 5 .O mEq/l) was present in 1l/74 ( 15%) of our arrest patients, seven of whom had a serum potassium above 7.0 mEq/l. In conclusion, we have confirmed that hypokalemia is frequently present in cardiac arrest. In at least half of the cases it may be associated with chronic diuretic administration unaccompanied by potassium supplementation. Our findings suggest that the hypokalemia is not caused by metabolic factors as suggested by Thompson and Cobb.’ The high incidence of vfib in association with hypokalemic arrests suggests that a low serum potassium level may be a trigger for lethal ventricular dysrhythmias in some patients. Although these results require confirmation by a prospective controlled study, our findings suggest that patients with chronically increased risk of vIib23 who receive long-term diuretic therapy should be checked periodically for hypokalemia and treated with chronic potassium supplementation if it develops. A patient with any of the following should be included in this recommendation: 1) definite ischemic heart disease, as indicated by definite or probable electrocardiographic (ECG) evidence of myocardial infarction, or definite angina pectoris; 2) definite or probable left ventricular hypertrophy on ECG; 3) chest-radiographic ray evidence of cardiac dilatation; or 4) definite clinical evidence of congestive heart failure. REFERENCES 1. Solomon RJ, Cole AG. Importance of potassium in patients with acute myocardial infarction. Acta Med Stand 1981;647:87-93. 2. Nordrehaug JE, Van der Lippe G. Hypokalemia and ventricular fibrillation in acute myocardial infarction. Br Heart J 1983;50:525-529. 3. Loewy EH. Hypokalemia and arrhythmias in myocardial infarction (letter). JAMA 1978;239:2113. 4. Martin D. Thiazide-induced hypokalemia: Association with acute myocardial infarction and ventricular fibrillation. JAMA 1978;239:43-45.

506

1985

5. Thompson RG, Cobb LA. Hypokalemia after resuscitation from out-of-hospital ventricular fibrillation. JAMA 1982;248:2860-2863. 6. Ornato JP, Fennigkoh L, Jaeger C. The electronic clipboard: An automated system for accurately recording events during a cardiac arrest. Ann Emerg Med 1981;10:138141. 7. Lillienfeld AM, Lillienfeld DE. Foundations of Epidemiology, 2nd edition. Oxford University Press, 1980:196. 8. Sterns RH, Cox M, Feig PU, et al. Internal potassium balance and the control of the plasma potassium concentration. Medicine 1981;60:339-354. 9. Kassirer JP, Harrington ST. Diuretics and potassium metabolism: A reassessment of the need, effectiveness, and safety of potassium therapy. Kidney Int 1977;11:505-515. 10. Edmonds CJ, Jasan 8. Total body potassium in hypertensive patients during prolonged diuretic therapy. Lancet 1972;2:8-12. 11. Fisch C. Relation of electrolyte disturbance to cardiac arrhythmias. Circulation 1973;47:408-419. 12. Holland OB, Nixon JV, Kuhnert L. Diuretic-induced ventricular ectopic activity. Am J Med 1981;70:762. 13. Papademetriou V, Fletcher R, Khatri IM, et al. Diuretic-induced hypokalemia in uncomplicated systemic hypertension: Effect of plasma potassium correction on cardiac arrhythmias, Am J Cardiol 1983;52:1017-1022. 14. Harrington ST, lsner JM, Kassirer JP. Our national obsession with potassium. Am J Med 1982;73:155-159. 15. Vlay SC, Reid PR. Ventricular ectopy: Etiology, evaluation, and therapy. Am J Med 1982;73:899-913. 16. Bigger JT. Definition of benign versus malignant ventricular arrhythmias: Targets for treatment. Am J Cardiol 1983;52:47c_54c. 17. Thomas RD. Ventricular fibrillation and initial plasma potassium in acute myocardial infarction. Postgrad Med J 1983;59:354-356. 18. Leemhuis MP, van Damme KS, Struyvenberg A. Effects of chlorthalidone on serum and total body potassium in hy pertensive patients. Acta Med Stand 1976;200:37-45. 19. Wilkinson PR, Hesp R, lssler H, et al. Total body and serum potassium during prolonged thiazide therapy for essential hypertension. Lancet 1975;1:759-762. 20. Wright BO, DiGiovanni AJ. Respiratory alkalosis, hypokalemia, and repeated ventricular fibrillation associated with mechanical ventilation. Anesth Analg 1969;49:467473. 21. Velebit V, Podrid P, Lown B, et al. Aggravation and provocation of ventricular arrhythmias by antiarrhythmic drugs. Circulation 1982;65:886-894. 22. Ruskin JN, McGovern 8, Garan H, et al. Antiarrhythmic drugs: A possible cause of out-of-hospital cardiac arrest. N Engl J Med 1983;309:1302-1306. 23. Hinkle LE. Short-term risk factors Acad Sci 1982;382:22-38.

for sudden

death. Ann NY