Misleading electrocardiographic results in patient with hyperkalaemia and diabetic ketoacidosis

Diabetic ketoacidosis is an important cause of morbidity and mortality in patients with insulin dependent diabetes. Around 2%-8% of all hospital admissions of diabetic patients are for ketoacidosis, and the overall mortality ranges from 2% to 10%.¹ Myocardial infarction is an uncommon but well recognised precipitating cause of diabetic ketoacidosis, accounting for 1% of cases. However, myocardial infarction and congestive cardiac failure account for 28% of deaths in diabetic ketoacidosis.² Initial baseline investigations of such patients include electrocardiography to assess possible precipitating causes and the effect of metabolic derangement on cardiac electrophysiology and rhythm.

Iatrogenic hypokalaemia is a major and avoidable cause of death in diabetic ketoacidosis. However, many diabetic patients with ketoacidosis present initially with hyperkalaemia, which may affect the electrocardiographic morphology. We present a case of a patient with diabetic ketoacidosis and hyperkalaemia whose initial electrocardiogram showed a pseudoinfarction pattern.

Case report

A 42 year old man with diabetes presented to accident and emergency with intermittent vomiting for three days. Over the previous two days he had been feeling increasingly tired, lethargic, polyuric, and thirsty. As he was unable to keep any food or fluids down, he had omitted two insulin injections, including one on the morning of admission. He had no history of chest pain. He had had type 1 diabetes for 15 years and was injecting soluble and isophane human insulin twice daily. He had not attended the hospital clinic for several years and had no known complications of diabetes. He was known to have hypercholesterolaemia but was not taking any drugs for it. He had a strong family history of ischaemic heart disease.

On initial assessment he was drowsy and dehydrated with tachypnoea, tachycardia, and a temperature of 37.1°C. There were no localising clinical signs of infection. A urine dipstick test detected ketones, bedside capillary testing with a glucometer showed “high” glucose concentrations, and arterial blood gas analysis indicated metabolic acidosis (pH 7.06, carbon dioxide partial pressure 1.8 kPa, oxygen partial pressure 15.9 kPa, bicarbonate concentration 4 mmol/l, and base excess –24 mmol/l). Blood samples were sent to the laboratory. We diagnosed diabetic ketoacidosis and started fluid replacement with 0.9% saline and intravenous soluble insulin at 6 units/hour.

An electrocardiogram showed sinus tachycardia and marked ST segment elevation in the anteroseptal leads consistent with anterior myocardial infarction (fig 1). No Q waves were present. In the inferior wall leads, there was ST segment elevation and T wave inversion in leads 3 and aVF. He also had an intraventricular conduction delay similar to a right bundle branch block with delayed rightward voltage. Chest radiography showed no abnormality. We provisionally diagnosed diabetic ketoacidosis, possibly precipitated by a silent acute myocardial infarction. The patient was transferred to the coronary care unit. In view of the metabolic derangement, thrombolytic therapy was withheld pending results of laboratory investigations.

Figure 1.

Electrocardiogram at presentation

Biochemical results showed the following serum concentrations: sodium 126 mmol/l, potassium 8.9 mmol/l, urea 23.1 mmol/l, creatinine 312 μmol/l, and glucose 83 mmol/l. He had a raised white blood cell count of 22.3×10⁹/l with neutrophils 87%. He was given 50 mmol/l intravenous sodium bicarbonate in addition to saline and insulin because of the severe hyperkalaemia and acidosis. Thrombolytic therapy was withheld until his hyperkalaemia was corrected. Repeat biochemical tests after two hours showed sodium 145 mmol/l, potassium 4.7 mmol/l and glucose 24 mmol/l. A repeat electrocardiogram (fig 2) showed sinus tachycardia with complete resolution of the anteroseptal and inferior ST elevation and the intraventricular conduction delay. He was therefore not given thrombolysis. His troponin I concentration 12 hours after admission was normal (0.1 μg/l). The patient subsequently made an uneventful recovery. His serum creatinine at discharge was 132 μmol/l.

Figure 2.

Electrocardiogram three hours after presentation

Discussion

Although total body potassium concentrations may be considerably depleted in diabetic ketoacidosis, plasma potassium concentrations at presentation are usually normal or high. Acidosis (which causes potassium ions to leave cells), insulin deficiency, and renal impairment all contribute to hyperkalaemia.² Potassium concentrations above 6.0 mmol/l have been reported in 22%-32% cases at presentation.³,⁴ Treatment with insulin and rising pH stimulate the entry of extracellular potassium into cells leading to a fall in extracellular concentrations.

Hyperkalaemia has profound effects on myocardial conduction and repolarisation and hence on the surface electrocardiogram. There is peaking of the T waves and sometimes shortening of the QT interval. The ST segment may virtually disappear, becoming incorporated into the proximal limb of the T wave. The P wave diminishes in amplitude progressively and eventually disappears when serum potassium concentrations are above 7.5 mmol/l. This may lead to a sinoventricular rhythm. Intraventricular conduction defect is manifested as widening of the QRS, which often resembles right bundle branch block with either a left anterior or a left posterior hemiblock. A sine wave pattern may occur in patients with end stage hyperkalaemia.⁵

Intraventricular conduction delay is well recognised in hyperkalaemia, but ST segment elevation or pseudoinfarction has been infrequently reported in diabetic ketoacidosis.^6–10 It is debatable whether the ST elevation is a primary repolarisation abnormality or an artefact caused by merging of the terminal R′ portion of the QRS with the T wave. It is also unclear whether the changes are due to acidosis or other metabolic abnormalities specific to diabetic ketoacidosis.¹⁰ Although an ischaemic basis for the electrocardiographic changes cannot be absolutely excluded, the rapid normalisation on resolution of hyperkalaemia and a normal troponin I concentration in our patient make ischaemia extremely unlikely.

This case shows that hyperkalaemia can simulate myocardial infarction and highlights the importance of having an efficient laboratory service in managing these critically ill patients. Myocardial infarction (which can be symptomatically silent in patients with long standing diabetes) is a well known precipitating factor of diabetic ketoacidosis.¹ With the current emphasis on reducing “door to needle” times for thrombolysis to curtail morbidity and mortality from coronary artery disease, it is worth remembering that metabolic abnormalities can sometimes alter the electrocardiographic appearances. Starting thrombolysis before metabolic abnormalities are corrected may therefore expose the patient to unnecessary drug treatment, along with its attendant risks.

Results of electrocardiography may mimic myocardial infarction in patients with severe metabolic disturbance