Discussion and Conclusions
Our study clearly demonstrated that the use of donepezil for more than a year significantly prolongs the PR, QRS and QT intervals. Male patients had a significantly greater prolongation of their QTc interval compared to females, and concomitant use of tricyclic antidepressants significantly increased the risk of QT interval prolongation, while no dose or treatment duration related differences were found.
The QT interval is an electrocardiographic representation of the time taken for depolarisation and repolarisation of the ventricles. Measuring the QT interval and understanding its significance can be difficult since it is influenced by multiple factors including heart rate and disease-associated ECG changes. The use of corrected QT intervals is more useful in the clinical setting, but the applicability of each formula remains debatable because of over or under correction at extremes of heart rate (Roden, 2008).
A prolonged QT interval has been used as a surrogate marker for predicting subsequent risk of ventricular arrhythmias. However, its significance varies between individuals and is affected by multiple factors including medications, underlying heart rhythm, asystolic pauses and genetics. Furthermore, the relationship between the QT interval and risk of TdP is non-linear. It should therefore be interpreted with caution and not in isolation, with consideration of the aforementioned confounding patient characteristics (Roden, 2008). Despite the above caveats, the risk of ventricular arrhythmias and increased cardiac morbidity and mortality in the presence of a prolonged QT interval are well established. Although its consequences may be unpredictable, QT prolongation does frequently lead to adverse outcomes (Kitt et al., 2015; Mandal et al., 2019; Roden, 2008; Malik & Batchvarov, 2000).
Our study confirmed that the use of donepezil for duration of at least 1 year was associated with a significant increase in the QT intervals. Donepezil enhances the cholinergic function of the brain while also causing undesired systemic cholinergic side-effects, with gastrointestinal disturbances such as nausea, vomiting and diarrhoea being the most common. The adverse effects on the heart are unpredictable. The heart is an organ rich in cholinesterase enzymes (ChE). Inhibition of these enzymes leads to increased acetylcholine, which deactivates the membrane bound voltage-gated calcium channels. Phase two of the cardiac action potential is also known as the plateau phase, which is defined by the membrane potential remaining constant due to the concurrent influx of calcium ions and efflux of potassium ions. Donepezil may delay this phase by increasing cardiomyocyte acetylcholine levels, hence deactivating the voltage-gated calcium channels, while simultaneously causing direct inhibition of delayed rectifier potassium channels. This results in QT interval prolongation and a subsequent increased risk of ventricular arrhythmias (Takaya et al., 2009; Malik & Batchvarov, 2000; Li & Cheng et al., 2017; Isik & Bozoglu et al., 2012).
Most studies to date have evaluated ECG changes in patients taking donepezil for up to four months. Significant prolongation of the PR interval is a common finding, with no associated change in the QT interval (Igeta et al., 2014; Wang et al., 2018; Rabkin & Cheng et al., 2016). We postulate that a longer duration of donepezil therapy may be necessary to demonstrate a significant QT interval prolongation, as found in our study. This finding raises concerns since the adverse effects of QT prolongation and subsequent TdP have only been previously identified in case studies, resulting in the risk being underestimated by clinicians (Kitt et al., 2015; Mandal et al., 2019). Furthermore, it is known that the QT interval increases with age and this may exacerbate the risk of ventricular arrhythmias in an older population (Rabkin et al., 2016). This risk may be further amplified by polypharmacy, electrolyte derangements and concurrent cardiac comorbidities or conduction disorders.
Female gender has been reported in literature to be an independent risk factor for QT prolongation, due to their greater baseline QT interval and oestrogen mediated gender differences in the expression of cardiac delayed rectifier potassium channels (Rabkin, 2015). Interestingly our study demonstrated that male patients experienced a significantly greater prolongation of their corrected QT interval compared to females. While females tend to show a longer QT interval at younger ages, gender differences decrease with age due to males having a more pronounced age-related increase in their QT interval (Rabkin, 2015). This is observed within our cohort with both older males and females having a similar QT interval prior to commencing donepezil therapy (Table 3).
A possible explanation is that male patients have a greater sensitivity to the increase in cardiac acetylcholine levels caused by donepezil, which results in a greater prolongation of the QT interval (Takaya et al., 2009). Naturally there is an increase in vagal tone and acetylcholine release when patients undergo a carotid sinus massage. Male patients exhibit an increased sensitivity to carotid sinus massage and have an increased prevalence of carotid sinus disease, which could theoretically be due to an underlying oversensitivity to acetylcholine (Tan & Newton et al., 2009; Kumar & Thomas et al., 2003; Kerr & Pearce et al., 2006). Further evidence of this is that male patients are less responsive to muscarinic receptor antagonists such as atropine (Tsutsui & Falcão et al., 2007). This mechanism may explain why male patients exhibit a greater prolongation of the corrected QT interval compared to females, but our results would require confirmation in a larger study.
Another manifestation of the increase in vagal tone caused by donepezil is prolonged PR and QRS intervals, as seen in our study population (Sato & Urbano et al., 2010). Underlying cardiac conduction disease and the concomitant use of commonly prescribed negatively chronotropic medications could exacerbate this further. We demonstrated that concurrent use of rate limiting calcium channel blockers and beta-blockers in those on donepezil therapy were associated with significant PR prolongation and a reduction in heart rate respectively. One patient in our study population developed symptomatic bradycardia requiring admission. Cessation of donepezil therapy reversed the bradycardia and resulted in normalisation of heart rate. Previous studies have similarly reported hospitalisation for donepezil associated bradycardia, with a proportion of patients developing syncope and irreversible heart block requiring permanent pacemaker implantation. Careful consideration must be taken when initiating donepezil in patients already prescribed negatively chronotropic medications (Bordier et al., 2003; Hernandez et al., 2009; Park-Wyllie & Mamdani et al., 2009).
Our study did not demonstrate a correlation between the duration of donepezil therapy and further QT prolongation (Table 5), but its prolonged use theoretically carries an increased risk of ventricular arrhythmias. The neurocognitive benefits of donepezil are most pronounced in the initial months of treatment and continue for up to two years. Following two years of treatment, the overall benefit of donepezil remains unclear (Courtney & Farrell et al., 2004). We noted that several of our patients had extended donepezil therapy for longer than the beneficial 2-year treatment period. We therefore suggest that such patients should be reassessed at this point and termination of treatment with donepezil should be considered if the risks of adverse events are deemed to outweigh any clear neurocognitive benefit.
We recognise the limitation of our study as a single centre study in an acute hospital setting with a predominantly Caucasian population, which may limit the generalizability of our results to other hospitals or community settings. We also recognise that ECG recordings took place during acute hospital admissions and the parameters measured may have been affected by confounding factors such as a concurrent acute illness e.g. infection. Our results were based on resting 12-lead ECGs before and during donepezil treatment, and despite providing valuable information they are only a snapshot of the patient’s cardiac conduction and may not be representative of their predominant rhythm. A Holter monitor would provide a more reliable assessment of cardiac conduction abnormalities and would also capture intermittent arrhythmias. Therefore a larger prospective study with prolonged cardiac monitoring and longer follow-up period would be of interest.
In conclusion, the results demonstrate that donepezil therapy results in a significant prolongation in the PR and QT intervals. Therefore, we urge clinicians to exercise caution and pharmacovigilance when prescribing donepezil, particularly in patients with pre-existing cardiac comorbidities or among those already prescribed medications known to cause QT prolongation such as tricyclic antidepressants. We suggest a resting 12-lead ECG should take place before and after donepezil initiation and where there is suspicion of QT prolongation, utilisation of a QT nomogram can reliably assess arrhythmogenic risk.