To our knowledge, this is the first study using an institutional dataset of pediatric patients to elucidate the relationship between potassium levels and risk of arrhythmia while on digoxin. There were four major findings in the study. First, consistent with our hypothesis, abnormal serum potassium level during digoxin administration was associated with increased odds of clinically significant arrythmia. Second, timing was significant as abnormal potassium level at the beginning of digoxin administration was not associated with arrhythmia. Third, the co-administration of drugs that interact with digoxin also increased odds of arrhythmia. Lastly, intravenous administration of digoxin was associated with an increased likelihood of experiencing an arrhythmia.
The increase in odds of arrhythmia due to abnormal potassium levels in this cohort is consistent with existing literature in adult patients. Arrhythmias are a documented side effect of digoxin primarily observed at supra-therapeutic concentrations of the drug.12 However, arrhythmias can be potentiated by hypokalemia, even at therapeutic or sub-therapeutic digoxin concentrations.13 From a pharmacodynamic perspective, digoxin competes with potassium ions for a binding site on the sodium-potassium ATPase. In a hypokalemic patient, there are potentially more binding sites available for digoxin, increasing the risk of unwanted side effects.14,15 Additionally, through a poorly characterized mechanism, digoxin increases parasympathetic tone which slows transmission through the conduction system of the heart and can cause bradyarrhythmia and ectopic rhythms.13,16 Additionally, low serum potassium on its own can slow atrioventricular node conduction such that the combination of digoxin and a hypokalemic state may exacerbate bradycardia.17 Our results are particularly important in light of the significant increase in the use of digoxin in pediatric patients with single ventricular congenital heart disease and those with heart failure, the large majority of which are at high risk for hypokalemia due to chronic diuretic therapy.
While previous literature primarily focuses on hypokalemia, in this study we considered both hypokalemia and hyperkalemia as independent risk factors. We did not get significant associations when we parsed patients into these groups. Additionally, experiencing both hypokalemia and hyperkalemia in the same encounter did not exacerbate risk. While it is possible that this result is due to insufficient power in the sub-analysis, it also suggests that the findings are not driven by one group and that both hypokalemic and hyperkalemic patients are at risk for arrhythmias. As a result, arrhythmia may be related to any kind of potassium instability, which can disrupt cardiomyocyte membrane potentials and catalyze aberrant conduction.18 Therefore, hypokalemia or hyperkalemia during a hospital encounter can be arrhythmogenic. In contrast to hypokalemia, hyperkalemia increases the risk of ventricular arrhythmias and when combined with bradycardia, it has been associated with an increase in fatality.19–21 In a study of pediatric patients presenting with digoxin poisoning, serum potassium and digoxin levels were elevated in those with recurrent arrhythmias compared to those who did not develop arrhythmias.22 This suggests vigilance is warranted for both hypokalemia and hyperkalemia.
The timing of abnormal potassium levels was also significant. Abnormal potassium levels before digoxin administration were not associated with arrhythmia. In contrast, abnormal potassium levels during the rest of the admission were associated with arrhythmia. While it is becoming more common to assess potassium levels prior to digoxin administration and provide replacement if indicated, this study suggests that one-time monitoring prior to the first dose of digoxin is inadequate. Paradoxically, the arrhythmia group was more likely to receive potassium replacement. This may reflect efforts to correct potassium imbalances prevalent in this group. However, increased repletion in the arrhythmia group did not appear to attenuate the association with arrhythmias, likely due to factors such as how promptly potassium was corrected, the rate of potassium infusion, repletion of other electrolytes, pH status, and use of medications that affect potassium distribution, such as beta-blockers or agonists.18
Drug-drug interactions can contribute to digoxin toxicity, as shown by the increase in odds of arrhythmia with each additional interacting drug. There are certain drug classes that demand extra caution when administered alongside digoxin, such as non-potassium sparing diuretics, certain antiarrhythmics, and beta adrenergic blockers.23 Increased toxicity in the presence of these drugs is due to different mechanisms, including increased serum digoxin concentration, potassium and magnesium wasting, and, as with beta-blockers, inducing an extracellular shift of potassium that leads to hyperkalemia.24 Further, there are numerous other drugs that can have impacts on digoxin concentration and metabolism.24,25 The administration of interacting drugs may not be avoidable because the majority of pediatric patients taking digoxin are also taking cardiac medications (96% in the non-arrhythmia group and 95% in the arrhythmia group). Polypharmacy, broadly defined as 5 or more medications taken daily, is associated with increased morbidity and mortality.25,26 Although pediatric patients on digoxin may require multiple drugs for therapy, frequent evaluations of drug regimens are necessary to minimize polypharmacy, especially in the setting of poorly controlled potassium levels. Not all drugs are equal, and some may be more arrhythmogenic than others when co-administered with digoxin.
Intravenous administration of digoxin was also associated with arrhythmia, highlighting the importance of drug formulation. There is already high variability in the ability of infants and children to clear digoxin compared to adults. Research on the pharmacokinetics of digoxin in children shows that this variability is further exacerbated after intravenous administration.23 This can lead to increased levels of serum digoxin and toxicity-related adverse events if not adequately monitored. In an inpatient setting, intravenous administration is often used for the loading dose, followed by oral digoxin for maintenance dosing. It is also the main administration route for patients who cannot take drugs orally.24 This suggests that careful administration is warranted for intravenous digoxin administration to lower the risk of arrhythmia.
While our study has identified important associations regarding serum potassium, digoxin administration, and arrhythmias, there are notable limitations. As a chart review, it relied on the completeness and accuracy of the electronic medical record. This is a single-center analysis, so the generalizability to other centers may be limited. Given the observational study design, we are unable to make causal claims between potassium levels and arrhythmias or account for unknown confounding variables. For example, if a patient had more than one encounter in the dataset, we treated them as two separate events for analysis. Given the relatively short half-life of digoxin and the lack of evidence that prior digoxin administration influences future responses to digoxin, we believe this was an appropriate methodological approach. However, there may be unpredictable interdependencies between encounters. Serum digoxin levels were very rarely obtained in any of the study patients, and we cannot comment the influence of digoxin level on our studied outcomes. We view the lack of measured digoxin levels as strengthening the salient point of our work. Our study focused on inpatient digoxin administration, and it may not necessarily translate to outpatient digoxin use. We did not assess outpatient events, such as arrhythmias or other adverse events that occurred between hospital encounters.