Hyperkalemia is one of the most common, potentially life-threatening metabolic disorders of reversible periarrest conditions that needs to be recognized and treated in time [9]. Potassium levels above the normal range often remain unnoticed and periarrest situations may occur without warning [10]. Electrocardiography is a widely used, easily attainable method to raise the possibility of hyperkalemia, however there have been conflicting reports about its sensitivity and specificity to signal elevated potassium levels [3, 5, 11].
In our study we investigated whether ECG alterations suggestive of hyperkalemia were present in a randomly selected group of normokalemic patients admitted to the Emergency Center. Previous studies have shown that the ability of physicians to predict hyperkalemia from the ECG was low with sensitivities between 0.43 and 0.34 and experienced readers’ ability to predict severity of hyperkalemia were likewise poor [5, 11]. ECGs have often been shown to be normal in hyperkalemia and a number of cases have been reported where patients with severely elevated potassium levels did not show typical ECG manifestations [7, 11]. In accordance with these investigations, we found that less than half of the hyperkalemic patients exhibited ECG changes suggestive of hyperkalemia, while the majority of the hyperkalemic patients showed no typical ECG changes at all. A surprisingly high proportion (24%) of normokalemic patients exhibited ECG alterations suggestive of hyperkalemia. Thus, based on ECG analysis alone, normokalemia and hyperkalemia cannot be confirmed or exluded in patients.
It must be noted, however, that some ECG changes suggestive of hyperkalemia (wide QRS, peaked-T waves, 1st degree AV-block and bradycardia) were significantly more prevalent in the severely hyperkalemic group. Although there wasn’t a significant difference between the frequency of ECG alterations suggesting hyperkalemia in normokalemic and moderately hyperkalemic patients, the number of ECG alterations suggestive of hyperkalemia simultaneously present increased with the degree of serum potassium elevation. These findings are in line with a previous report on the higher frequency of ECG changes with increasing potassium levels [4]. Peaked T waves are considered to be the typical earliest ECG signs of elevated serum potassium levels [3,4,5]. In our study, peaked T-waves was the second most common ECG manifestation among severely hyperkalemic patients, while wide QRS was the most common ECG change and significantly more often found among all hyperkalemic patients compared to patients with normal potassium levels.
We examined the prevalence of four ECG abnormalities, whose association with hyperkalemia has been found to be equivocal according to previous studies. ST depression may be an ECG manifestation of hyperkalemia [12], but we did not detect a significant difference in the frequency of ST depression between normokalemic and hyperkalemic patients in our study. Atrial fibrillation has been associated with certain changes in the ECG and most studies have found that lower potassium levels were associated with a higher risk of atrial fibrillation [13,14,15]. Our results showed, however, that atrial fibrillation was more prevalent in severely hyperkalemic patients compared to normokalemic patients. We attribute these results to the synergistic effect of two groups of diseases often present in patients with high potassium levels. Hyperkalemia and heart failure are common in chronic kidney disease and heart failure is often the cause of or caused by atrial fibrillation. Therefore, atrial fibrillation occurs not as the result of hyperkalemia but rather as the consequence of illnesses often associated with hyperkalemia.
Although shortening of the QTc interval in hyperkalemia has been reported in several investigations [12, 16], the possibility of prolonged QTc occurance in hyperkalemia has also been raised [17]. In our study, prolonged QTc were more frequent in severely hyperkalemic patients compared to normokalemic patients. In fact, prolonged QTc was the only ECG alteration significantly more frequent in both moderately and severely hyperkalemic patients compared to the group with normal potassium levels. Although the reasons underlying our findings need to be clarified, the results imply that prolonged QTc and atrial fibrillation could also draw attention to hyperkalemia, besides the more acknowledged ECG manifestations of hyperkalemia.
Our study has other implications. When we compared baseline data from patients in the normokalemic and hyperkalemic groups, CKD and comborbidities such as heart and liver failure were significantly more frequent in hyperkalemic patients. CKD and/or hemodialysis have been known to increase the likelihood of hyperkalemia [4, 18] and so patients suffering from renal failure are more at risk of developing potassium cardiotoxicity [19]. The usage of certain types of medication, including ACEI, ARB and potassium-sparing diuretics has been associated with an increased number of hyperkalemia-related hospitalizations and mortality [20,21,22,23]. In keeping with earlier studies, we found that the application of these types of drugs was also more common in patients with elevated potassium levels. Our data underline the importance of regular monitoring of electrolytes in patients taking hyperkalemia-inducing medication, preferably already in non-urgent situations within the primary care setting since ECG diagnosis of hyperkalemia is uncertain.
To our knowledge, this study is the first to investigate ECG alterations suggestive of hyperkalemia in a large number of patients with normal potassium levels. Besides supporting evidence for the unreliability of ECG diagnosis in hyperkalemia, our results show that a fourth of normokalemic patients also exhibit ECG alterations suggestive of hyperkalemia. This indicates that treatment of suspected hyperkalemia in the prehospital setting, prior to laboratory confirmation of potassium levels, may not be prudent. Although it has been suggested that initiation of life-saving treatment with calcium in suspected hyperkalemia prior to laboratory confirmation of hyperkalemia would be advisable, we disagree with this proposition [24]. Investigations have shown, that the empiric treatment of hyperkalemia based on ECG alone was predicted to lead to the mistreatment of at least 15% of patients [11] and treatment decisions should not depend only on the presence of ECG alterations [4].
Limitations
Our study has several limitations. The normokalemic (control) group and hyperkalemic group were not matched regarding age, medication and comorbidites. This was to be expected, however, as this investigation was a cross-sectional study of patients admitted to emergency care. The interpretation of ECGs may be another confounding factor, since a number of ECG alterations can be due to other causes than hyperkalemia or alternatively, ECG changes due to other conditions may mask signs of hyperkalemia. We argue, however, that in the prehospital, emergency setting information regarding medication and previous illnesses is not readily available to the caregiver and therefore, he or she must make decisions regarding diagnosis and treatment with none or very limited information. Another limitation of our study is that although the investigation was conducted on a relatively large number of patients, a larger-scale study examining the ECG alterations suggestive of hyperkalemia in normokalemic patients would be needed to confirm our results.