Admission serum potassium in patients with acute myocardial infarction - Its correlates and value as a determinant of in-hospital outcome

Study objectives: Although controversial, hypokalemia (LK) in patients with acute myocardial infarction (MI) is thought to predict increased in-hospit al morbidity, particularly cardiac arrhythmias, and mortality. Also, the me chanism of low serum potassium in the setting of MI has not been delineated . We evaluated the frequency, attributes, and outcome, and speculated on th e mechanism of LK in patients with MI. Design: This was a prospective cross-sectional study of 517 consecutive pat ients with MI admitted to the coronary care unit (CCU). Serum potassium was measured in the emergency department and repeatedly thereafter throughout hospitalization, and was used in the analysis, along with a large array of clinical and laboratory variables. Results: The patients were allocated to a LK and a normokalemic (NK) cohort , based on the emergency department serum potassium measurement. The 41 pat ients with LK (3.16 +/- 0.24 mEq/L; 7.9% of total) were comparable on admis sion in their baseline assessment to the 476 patients with normal serum pot assium (4.28 +/- 0.56 mEq/L), except for lower emergency department magnesi um (1.48 +/- 0.15 mg/dL vs 1.96 +/- 0.26 mg/dL; p = 0.0005) and earlier pre sentation after onset of symptoms (3.0 +/- 4.1 h vs 1.4 +/- 6.2 h; p = 0.05 ). There was a poor correlation between serum potassium and magnesium on ad mission (r = 0.14). Peak creatine kinase (CK) and myocardial isomer of CK w ere higher in the LR patients (3,870 +/- 3,810 IU/L vs 2,359 +/- 2,653 IU/L [p = 0.018] and 358 +/- 312 IU/L vs 228 +/- 258 IU/L [p 0.013], respective ly). Management of the two cohorts was the same, except for a higher rate o f use of magnesium (14.6% vs 4.6%; p = 0.007), serum potassium supplements (80.2% vs 43.1%; p = 0.000005), and antiarrhythmic drugs (78.0% vs 50.4%; p = 0.0007) in the LK patients. No difference was detected between the LK an d NK patients in total mortality (24.4% vs 18.3%; p = 0.31), cardiac mortal ity (17.1% vs 15.3%; p = 0.52), atrial fibrillation (14.6% vs 13.9%; p = 0. 89), and ventricular tachycardia (22.0% vs 16.0%; p = 0.32), but ventricula r fibrillation (VF) occurred more often (24.4% vs 13.0%; p = 0.04) in the L K patients. However, proportions of VF occurring in the emergency departmen t, CCU, or wards in the two cohorts were not different, but they were highe r during the time interval prior to emergency department admission in LK pa tients (17.1% vs 2.1%; p = 0.00001). Conclusions: LK is seen in approximately 8% of patients with MI in the emer gency department; LK is associated with low emergency department magnesium, and low serum potassium levels in the CCU and throughout hospitalization. LK has no relationship to preadmission use of diuretics, it is associated w ith early presentation to the emergency department, and it is not a predict or of increased morbidity or mortality.