Hypokalaemia and ventricular fibrillation in acute myocardial infarction

Serum potassium concentrations obtained on admission to hospital were inversely related to the incidence of ventricular fibrillation in 289 women and 785 men with acute myocardial infarction, 92 of whom developed ventricular fibrillation. Hypokalaemia (serum potassium concentration less than or equal to 3.5 mmol/l) was found in 122 patients (11.4%). The incidence of ventricular fibrillation was significantly greater in patients with hypokalaemia compared with those classified as normokalaemic (serum potassium concentration greater than or equal to 3.6 mmol/l) (17.2% v 7.4%). The increased risk of ventricular fibrillation in the hypokalaemic group was about the same for women and men. While they were in hospital patients with hypokalaemia developed ventricular fibrillation significantly earlier than did normokalaemic patients (median 0.3 hours v 7 hours). Hypokalaemia was more common in women (17.3%) than in men (9.2%), and 55% of the hypokalaemic patients had been treated with diuretics before admission compared with 22% of the normokalaemic group. Hypokalaemia on admission to hospital predicts an increased likelihood and early occurrence of ventricular fibrillation in patients with acute myocardial infarction.     

Effect of timolol on changes in serum potassium concentration during acute myocardial infarction.

One hundred and six patients with acute myocardial infarction admitted to hospital within four hours after the onset of symptoms were randomised to treatment with intravenous timolol (54 patients) or placebo (52 patients). Serum potassium concentrations were estimated at frequent intervals during the first 24 hours of admission. Patients in both treatment groups, who did not receive subsequent diuretic treatment, had a transient rise in serum potassium concentration, which was maximal after four hours. This rise was abolished by diuretic treatment in the placebo group but not in the timolol group, in which there was a pronounced and prolonged rise in serum potassium concentration. The change in serum potassium concentration in the first four hours after admission correlated with cumulative creatine kinase release in the placebo group, but not in the timolol group. Hypokalaemia (serum potassium concentration less than or equal to 3.5 mmol/l) occurred in 15 (28.8%) patients in the placebo group and in seven (13%) in the timolol group and was independent of infarct size. The frequency of hyperkalaemia was not increased in the timolol group. By increasing the serum potassium concentration and preventing hypokalaemia, the use of intravenous timolol early in acute myocardial infarction may have important clinical effects in addition to reducing infarct size.     

Effect of timolol on changes in serum potassium concentration during acute myocardial infarction

One hundred and six patients with acute myocardial infarction admitted to hospital within four hours after the onset of symptoms were randomised to treatment with intravenous timolol (54 patients) or placebo (52 patients). Serum potassium concentrations were estimated at frequent intervals during the first 24 hours of admission. Patients in both treatment groups, who did not receive subsequent diuretic treatment, had a transient rise in serum potassium concentration, which was maximal after four hours. This rise was abolished by diuretic treatment in the placebo group but not in the timolol group, in which there was a pronounced and prolonged rise in serum potassium concentration. The change in serum potassium concentration in the first four hours after admission correlated with cumulative creatine kinase release in the placebo group, but not in the timolol group. Hypokalaemia (serum potassium concentration less than or equal to 3.5 mmol/l) occurred in 15 (28.8%) patients in the placebo group and in seven (13%) in the timolol group and was independent of infarct size. The frequency of hyperkalaemia was not increased in the timolol group. By increasing the serum potassium concentration and preventing hypokalaemia, the use of intravenous timolol early in acute myocardial infarction may have important clinical effects in addition to reducing infarct size.     

Hypokalemia,Arrhythmias and Early Prognosis in Acute Myocardial Infarction

ABSTRACT Serum potassium concentration was estimated on admission to hospital in 289 women and 785 men with acute myocardial infarction. The proportion of women in potassium subgroups was inversely related to serum potassium concentration, increasing from 8% at serum potassium ≥5.2 mmol/1 to 58% at ≤3 mmol/1. The frequency of diuretic therapy was also higher in women (35%) than in men (23%). The mortality rate was high at 3 months in patients with one or more arrhythmias (atrioventricular block grade 2, complete heart block, bundle branch block, atrial fibrillation, premature ventricular contractions, ventricular tachycardia) detected by conventional methods during the first 48 hours after admission. Hypokalemia (serum potassium ≤3.5 mmol/1) did not significantly predict increased occurrence of any of these arrhythmias. Small inhomogeneities of arrhythmias between the potassium groups may have been caused by digitalis therapy prior to admission. Hypokalemia on admission did not predict altered prognosis during the first 3 months.     

Cardiac arrhythmias following acute myocardial infarction: Associations with the serum potassium level and prior diuretic therapy

The relationship between the initial serum potassium level andthe incidence of cardiac arrhythmias following myocardial infarctionhas been reviewed in a coronary care unit setting. The incidenceof arrhythmias in general, and ventricular fibrillation, ventriculartachycardia and frequent ventricular eclopic beats in particular,were inversely related to the initial serum potassium level.Hyperkalaemia was also significantly associated with ventricularfibrillation and ventricular tachycardia. Hypokalaemia was significantlymore common in patients previously treated with diuretics, thoughmost patients with hypokalaemia had not been so treated. Theoccurrence of an acute hypokalaemic syndrome, independent of,but exacerbated by, diuretic therapy, is further supported bythese results.     

Effect of intravenous adrenaline on electrocardiogram, blood pressure, and serum potassium.

Increased catecholamines after myocardial infarction may contribute to the development of arrhythmias. We have infused adrenaline intravenously in nine normal volunteers to levels similar to those seen after myocardial infarction. Adrenaline caused an increase in systolic blood pressure, a decrease in diastolic blood pressure, and an increase in heart rate. Adrenaline also produced a decrease in T wave amplitude and an increase in the QTc interval. The serum potassium fell dramatically during the adrenaline infusion from a control value of 4.06 mmol/l to 3.22 mmol/l. Hypokalaemia after myocardial infarction is associated with an increased incidence of ventricular arrhythmias. Thus, circulating adrenaline may increase the frequency of arrhythmias both directly via changes in ventricular repolarisation and indirectly via adrenaline induced hypokalaemia.     

Effect of intravenous adrenaline on electrocardiogram, blood pressure, and serum potassium

Increased catecholamines after myocardial infarction may contribute to the development of arrhythmias. We have infused adrenaline intravenously in nine normal volunteers to levels similar to those seen after myocardial infarction. Adrenaline caused an increase in systolic blood pressure, a decrease in diastolic blood pressure, and an increase in heart rate. Adrenaline also produced a decrease in T wave amplitude and an increase in the QTc interval. The serum potassium fell dramatically during the adrenaline infusion from a control value of 4.06 mmol/l to 3.22 mmol/l. Hypokalaemia after myocardial infarction is associated with an increased incidence of ventricular arrhythmias. Thus, circulating adrenaline may increase the frequency of arrhythmias both directly via changes in ventricular repolarisation and indirectly via adrenaline induced hypokalaemia.     

低钾血症对急性心肌梗死患者预后的影响

目的探讨急性心肌梗死(AMI)患者低钾血症的发生情况及其对预后的影响。方法对929例ST段抬高的AMI患者于入院时抽血测定血钾、肌酸激酶(CK)、肌酸激酶同工酶(CK-MB)、肌钙蛋白I(cTnI),根据血钾水平分为低血钾组(血钾<3.5 mmol/L)和正常血钾组(血钾3.5~5.5 mmol/L),同时观察住院期间严重不良事件(室性心动过速、心室颤动和猝死)的发生情况。结果低钾血症的发生率为13.7%,下壁+后壁AMI的发生率最低(10.4%),明显低于下后壁+右心室、前间壁和广泛前壁心肌梗死;发病至抽血时间≤3 h的低钾血症发生率为17.3%,明显高于发病时间>3 h者;低血钾组的CK、CK-MB和cTnI峰值明显高于正常血钾组;低血钾组总的严重不良事件发生率(23.8%)明显高于正常血钾组(15.8%)。结论低钾血症与AMI患者的梗死时间、部位和面积相关,并严重影响患者的预后。     

Ventricular fibrillation in the initial phase of acute myocardial infarct

In order to determinate the incidence, predictivity and prognosis of ventricular fibrillation in the early phase of acute myocardial infarction a series of 301 patients with acute myocardial infarction consecutively assisted by the Mobile Coronary Care Unit of Florence was analyzed. 151 patients (50.2%) received intensive care within 2 hours from the onset of the symptoms, 75 patients (24.9%) received intensive care between the second and sixth hour. 38 patients (12.6%) had at least one episode of ventricular fibrillation. 30% of the episodes of ventricular fibrillation happened within 1 hours from the onset of the symptoms, 47.4% within 2 hours, 74% within 6 hours. Serious arrhythmias complicated the early phase of acute myocardial infarction, but only sinus bradycardia seems to have a significant predicativity of ventricular fibrillation (P less than 0.05). We found that hospital survival resuscitated patients is strictly related to the time between early symptoms and the episode of ventricular fibrillation: 91% of the patients with ventricular fibrillation within 1 hour were discharged alive from hospital, 71% of those with ventricular fibrillation within 6 hours, 20% of those with ventricular fibrillation beyond 6 hours (P less than 0.01). The high rate and the favourable prognosis of ventricular fibrillation in the early phase of acute myocardial infarction must lead to a widespread implementation of rapid response emergency care systems away from hospital.     

The role of plasma renin activity in evaluating the adequacy of mineralocorticoid replacement in primary adrenal insufficiency

BACKGROUND Elevation of plasma renin activity (PRA) is a feature of mineralocorticoid deficiency in patients with primary adrenal insufficiency. This study was designed to assess the usefulness of PRA as an index of adequacy of fludrocortisone (FC) replacement in patients with primary adrenal failure, paying particular attention to the variability in PRA levels during FC and glucocorticoid treatment. METHODS Twenty-two patients with mineralocorticoid deficiency due to primary adrenal disease were studied at 3 time points: 8, 24 and 32 hours following the administration of FC replacement. Body weight, blood pressure while supine and erect, PRA, and plasma or serum levels of aldosterone, urea, sodium and potassium were measured at each time. The clinical and biochemical consequences of adjusting the FC dose were monitored in 5 patients with PRA levels above the range seen in normal subjects and in one hypokalaemic patient with normal PRA levels. RESULTS At 8 and 32 hours following FC administration, PRA levels were not significantly different. PRA levels were significantly higher at 32 hours following FC administration (4.7 ±1.1 nmol/l/h) than at 24 hours (4.2 ±1.1 nmol/l/h, mean ± SEM, P <0.05). At 8 and 32 hours following FC administration, potassium levels were similar. Potassium levels were significantly higher at 32 hours following FC administration (3.9 ± 0.1 mmol/l) than at 24 hours (3.6 ±0.1 mmol/l, P < 0.05). No changes in measurements of sodium, urea, mean supine and erect arterial pressure or body mass index were noted at the different study points. Attempted lowering of elevated PRA in 5 normokalaemic subjects by raising the dose of FC led to normalization of PRA in all of these patients but 2 developed hypokalaemia and oedema. Lowering of FC dose in one hypokalaemic patient with normal PRA levels led to the PRA levels rising to a supranormal value while the hypokalaemia was corrected. CONCLUSIONS These results indicate that when plasma renin activity is estimated in patients with primary adrenal insufficiency replaced with daily doses of fludrocortisone, the time of day of blood sampling is not critical. Lowering elevated plasma renin activity levels to normal in patients who were considered to be otherwise normal may lead to over-treatment in some patients. Therefore, optimal fludrocortisone replacement may be associated with mildly elevated plasma renin activity levels. The information obtained by monitoring plasma renin activity adds little to the assessment of patients based on clinical evaluation and measurement of urea and electrolyte levels in blood.     

Late clinical outcome in patients with early ventricular fibrillation after myocardial infarction

Whether ventricular fibrillation occurring within 48 h after acute myocardial infarction is associated with particular clinical features and poor prognosis, especially in patients with anterior myocardial infarction, is still debated. Therefore, clinical variables and in-hospital and 1 year mortality rates were analyzed in 2,088 patients, aged 18 to 95 years (mean +/- SD 64 +/- 12), admitted to the hospital with acute myocardial infarction between 1979 and mid 1984. One hundred forty-seven patients (7%) had at least one episode of ventricular fibrillation occurring within 48 h of hospital admission. Of these, 25% died during their initial hospitalization compared with 13% of patients without early ventricular fibrillation (p less than 0.001). In greater than 50% of patients with early ventricular fibrillation, the immediate cause of death was left ventricular failure or cardiogenic shock. In contrast, the 1 year mortality rate after hospital discharge was not significantly greater in patients with than in those without early ventricular fibrillation (15 versus 11%, respectively), particularly in the subgroup of patients with anterior myocardial infarction in which the mortality rate tended to be lower in patients with early ventricular fibrillation (8 versus 14%, respectively). Similar mortality results were found when only primary (not associated with left ventricular failure) ventricular fibrillation was analyzed. The left ventricular ejection fraction and the incidence of complex ventricular arrhythmias from 24 h ambulatory electrocardiographic monitoring obtained at hospital discharge were not different in survivors with or without early ventricular fibrillation (0.45 +/- 0.13 versus 0.49 +/- 0.14 and 41 versus 41%, respectively).(ABSTRACT TRUNCATED AT 250 WORDS)     

Early use of metoprolol and serum potassium in suspected acute myocardial infarction

In 1350 patients with suspected acute myocardial infarction, serum potassium was analysed in the emergency ward. The effect of metoprolol was compared with placebo in a double-blind randomized trial. Metoprolol increased serum potassium from 4.11 +/- 0.02 mmol/l to 4.27 +/- 0.02 mmol/l (P less than 0.001) during the 1st day after hospital admission, whereas serum potassium levels remained fairly constant in patients given placebo during the same time (4.11 +/- 0.02 to 4.14 +/- 0.02 mmol/l; P greater than 0.2). Similar results were obtained when analysing patients with a confirmed myocardial infarction separately. The effects appeared homogeneously distributed in subgroups related to sex, clinical history, infarct site, infarct size and delay time from onset of symptoms to start of treatment. We conclude that early treatment with the beta-1-selective blocker metoprolol in patients with suspected acute myocardial infarction increases serum potassium.     

Polymorphic ventricular tachycardia in acute myocardial infarction without ST elevation in a patient with thrombocytopenia

67-year-old woman with thrombocytopenia (treated with prednisolon and azathiopryn) was admitted because of acute myocardial infarction without ST segment elevation (NSTEMI). From the 2nd day we observed increasing QTc interval from 461 ms with normal potassium level. Suddenly on the 6th day of the so far uncomplicated AMI ventricular fibrillation developed and was successfully treated with DC shock, and amiodarone (150 mg i.v.) was administered because of recurrent NSVT. Potassium level was 2.9 mmol/l. Within the next 2 days in the morning hours we observed episodes of recurrent polymorphic ventricular tachycardia (PMVT), always progressing into ventricular fibrillation (VF). The ECG showed QT interval--520 ms, QTc--602 ms. The patient was given an increasing dose of beta-blocker and lidokaine in i.v. infusion. After this regimen PMVT/VF did not recur and QT was normalized. Additionally successful PCI of LAD with 80% stenosis was performed. The paper discusses the problem of PMVT in the settings of AMI.     

Confirmatory testing in normokalaemic primary aldosteronism: the value of the saline infusion test and urinary aldosterone metabolites

OBJECTIVE: Primary aldosteronism has recently been recognized as the most frequent cause of secondary hypertension. Since most patients are normokalaemic, differentiation to essential hypertension is challenging. As differentiation by baseline aldosterone/renin ratio may be insufficient, diagnosis should be confirmed by additional tests. However, as most confirmatory tests have been evaluated in hypokalaemic primary aldosteronism only, we reassessed the value of the saline infusion test and 24 h urinary aldosterone metabolites as confirmatory tests for both normo- and hypokalaemic primary aldosteronism under current antihypertensive medication. PATIENTS AND METHODS: 25 patients with primary aldosteronism (11 hypokalaemic, 14 normokalaemic), 29 patients with essential hypertension and 47 normotensive subjects were studied. The hypertensives received their usual medication with the exception of spironolactone. All subjects underwent a standard saline infusion test (determination of plasma aldosterone before and after 2.0 liters of isotonic saline for 4 hours i.v.) and collected a 24 h urine sample for examination of urinary tetrahydroaldosterone and aldosterone-18-glucuronide. RESULTS: In hypokalaemic primary aldosteronism the saline infusion test showed a reasonable sensitivity (91%) and specificity (90%). However, the test failed to differentiate sufficiently between essential hypertension and normokalaemic primary aldosteronism (sensitivity 57%, specificity 90%). Similarly, urinary tetrahydroaldosterone had higher sensitivity in hypokalaemic than in normokalaemic primary aldosteronism (sensitivity 64% vs 36%, specificity 100%), whereas for aldosterone-18-glucuronide, no differences in hypo- and normokalaemic primary aldosteronism were found (sensitivity 45% and 43%, specificity 100%). CONCLUSIONS: These data show that the saline infusion test as an established test in classical hypokalaemic primary aldosteronism is not a reliable test in the normokalaemic variant of the disease. Due to its low accuracy, determination of urinary aldosterone metabolites did not prove useful in confirming either normo- or hypokalaemic patients. We conclude from our data that these tests should not be used as confirmatory testing in the normokalaemic variant of primary aldosteronism.     

Influence of increased adrenergic activity and magnesium depletion on cardiac rhythm in alcohol withdrawal.

OBJECTIVE--To investigate the prevalence of arrhythmias in alcoholic men during detoxification and its relation to neuroendocrine activation and electrolyte disturbances. DESIGN--Consecutive case-control study. SETTING--Primary and secondary care, detoxification ward. PATIENTS AND CONTROLS--19 otherwise healthy alcoholic men (DSM-III-R) with withdrawal symptoms necessitating detoxification in hospital. 19 age matched, healthy non-alcoholic men as controls for Holter recordings. INTERVENTIONS--Treatment with chlomethiazole; additional treatment with carbamazepine in patients with previous seizures. MAIN OUTCOME MEASURES--Computer based analyses of mean heart rate and arrhythmias from 24 hour Holter recordings, 24 hour urinary excretion of adrenaline and noradrenaline, magnesium retention measured by means of intravenous loading test, and serum concentrations of electrolytes. RESULTS--The 24 hour mean heart rate was higher in the alcoholic men (97.4 beats/minute, 95% confidence interval (CI) 91.2 to 103.6) than in the controls (69.6 beats/minute, 95% CI 65.4 to 73.8, P < 0.001). However, there was no difference in diurnal heart rate variation. The prevalence of premature supraventricular depolarisations was lower in the alcoholic men (P < 0.05). Neither atrial fibrillation nor malignant ventricular arrhythmias occurred. The sinus tachycardia in the alcoholic men correlated with the concomitant urinary excretion of catecholamines (P < 0.05). The mean serum magnesium concentration was 0.78 mmol/l (95% CI 0.73 to 0.83) in the alcoholic men and 0.83 mmol/l (95% CI 0.81 to 0.85) in a reference population of 55 men aged 40. Magnesium depletion (defined as magnesium retention > 30%) was detected in 10 alcoholic men (53%). Three alcoholic men had serum potassium concentrations < or = 3.3 mmol/l on admission. CONCLUSION--Increased adrenergic activity, magnesium depletion, and hypokalaemia are often seen after heavy drinking, but in alcoholic men without clinical heart disease these changes were not accompanied by arrhythmias other than sinus tachycardia during detoxification in hospital.     

Prognostic implications of ventricular fibrillation in acute myocardial infarction: new strategies required for further mortality reduction

OBJECTIVE—To determine the changing risk of ventricular fibrillation, the prognostic implications, and the potential long term prognostic benefit of earlier hospital admission, after acute myocardial infarction.
DESIGN—Prospective observational study.
SETTING—A district general hospital in east London.
PATIENTS—1225 consecutive patients admitted to a coronary care unit with acute myocardial infarction.
MAIN OUTCOME MEASURES—Time of onset of pain and ventricular fibrillation, and long term survival of patients admitted with acute myocardial infarction.
RESULTS—The rate of ventricular fibrillation in these hospital inpatients was high in the first hour from onset of pain (118 events/1000 persons/h; 95% confidence interval (CI) 50.7 to 231) and fell rapidly to an almost constant low level by six hours; 27.4% of patients with early ventricular fibrillation died in hospital, compared with 11.6% of those without (p < 0.0001), but mortality in patients who survived to hospital discharge was not altered by early ventricular fibrillation (five year survival: 75.0% (95% CI 60.0% to 84.8%) with ventricular fibrillation v 73.3% (95% CI 69.6% to 76.6%) without ventricular fibrillation).
CONCLUSIONS—Patients successfully resuscitated from early ventricular fibrillation have the same prognosis as those without ventricular fibrillation after acute myocardial infarction. Faster access to facilities for resuscitation must be achieved if major improvements in the persistently high case fatality of patients after acute myocardial infarction are to be made.


Keywords: ventricular fibrillation; acute myocardial infarction; prognosis     

Electrocardiographic changes in thyrotoxic periodic paralysis.

The electrocardiograms (ECGs) of 30 patients with hypokalaemic thyrotoxic periodic paralysis during and after paralysis were studied. During paralysis, typical features of hypokalaemia were seen in all patients with serum potassium levels of 2.8 mmol/l or less; above this level, the ECGs varied from non-diagnostic to those showing typical features of hypokalaemia. It was not possible to accurately predict the serum potassium level from the ECG except when either sinus arrest or heart block was present. Although extrasystoles have been reported to be common in hypokalaemia, none of the patients in this study had extrasystoles. Sinus arrest occurred in two patients and second degree atrio-ventricular block occurred in three patients, a finding which has not been reported in hypokalaemia.     

Mechanism of myocardial injury in dogs with hypokalaemia

Hypokalaemia was induced by infusing polystyrene sulphonate into the colon of mongrel dogs. Sixty minutes after infusion adrenaline 10 micrograms.kg-1 was injected intravenously, which had no effect on serum creatine kinase activity or myocardial histology in the control dogs. However, in dogs with hypokalaemia creatine kinase activity was increased, and pronounced histological changes were seen 60 min after injection. A clear reciprocal relation was found between serum potassium concentration and creatine kinase activity. Premedication with an alpha 1 blocking agent prevented the changes associated with hypokalaemia. Heart mitochondria were prepared from other dogs with hypokalaemia 5 min after adrenaline injection and their calcium content measured. Heart mitochondrial calcium content was increased in parallel with the decrease in serum potassium concentration. Alpha 1 blockade also prevented the increase in mitochondrial calcium content. These results indicate that the intracellular calcium concentration is considerably increased by alpha 1 receptor stimulus under hypokalaemic conditions and that this increase in calcium concentration plays a crucial role in the genesis of myocardial damage. Since adrenaline increases coronary blood flow small doses of adrenaline in subjects with hypokalaemia may lead to the development of myocardial injury not associated with ischaemia.     

Ouabain binding and inotropy in acute potassium depletion in guinea pigs

In hypokalaemia the incidence of cardiac toxicity with digitalis is increased, possibly through changes in the affinity or capacity of the digitalis receptor in the heart. Previous studies have reported an increased ouabain binding capacity or Na+-K+ ATPase activity after hypokalaemia in human erythrocytes and rabbit and guinea pig hearts and no change or decreases in rabbit or rat skeletal muscles with no changes in ouabain affinity. The present study determined (a) the effect of potassium on 3H-ouabain binding to normokalaemic guinea pig cardiac cell membranes, (b) the effect of acute hypokalaemia induced by a potassium deficient diet for 14-18 days in guinea pigs on 3H-ouabain binding to erythrocytes and cardiac and skeletal muscle homogenates, and (c) ouabain induced inotropy in isolated contracting guinea pig left atria, right ventricular papillary muscles, and soleus muscle strips from normokalaemic and hypokalaemic guinea pigs. 3H-ouabain binds to cardiac cell membranes in the presence of magnesium and inorganic phosphate with an affinity (KD value) of 1.13 X 10(-7) mol X litre-1. Potassium decreased this affinity without changing the binding capacity. Erythrocytes and heart muscle homogenates showed the same affinity as cardiac cell membranes, whereas soleus muscle homogenates had a higher affinity for ouabain (KD 5.1 X 10(-8) mol X litre-1). After hypokalaemia, the ouabain affinity did not change in any tissue.(ABSTRACT TRUNCATED AT 250 WORDS)     

Arrhythmogenic potential of diuretic induced hypokalaemia in patients with mild hypertension and ischaemic heart disease.

In view of evidence suggesting an association of mild hypokalaemia with cardiac arrhythmia, the arrhythmogenic potentials of potassium losing and potassium sparing diuretic treatments were compared in a controlled prospective crossover study of 10 patients with mild hypertension and ischaemic heart disease. Mean (SEM) plasma potassium was 4.3(0.06) mmol/l and 3.3(0.07) mmol/l after potassium sparing and potassium losing treatments respectively. Blood pressure and volume depletion as assessed by weight change, plasma renin activity, and noradrenaline concentrations did not differ significantly in the two treatment periods. The potassium losing treatment phase was associated with an increased frequency of ventricular extrasystoles, a higher Lown grading during ambulatory electrocardiographic monitoring, prolonged duration and decreased phase 0 velocity of the monophasic action potential, a prolonged ventricular effective refractory period, and increased myocardial electrical instability as assessed by programmed ventricular stimulation. It is concluded that minor changes in plasma potassium concentration are associated with increased ventricular electrical instability in patients with ischaemic heart disease. Mild hypokalaemia in such patients may predispose to life threatening arrhythmias and should be avoided.