The pathophysiology of conversion

ECT appears to be associated both with cardioversion from AF to sinus rhythm as well as the new development of AF in patients with baseline sinus rhythm [7,8,9,10,11,12,13,14,15,16,17,18,19,20]. However, the vectors used in ECT (across the brain) differ significantly from those used in synchronized cardioversion for AF (across the heart). In fact, current studies of the electrical field generated by ECT remain poorly defined but do not appear to include the body beyond the brain [21]. This suggests that there may be different pathophysiology for rhythm changes during or after ECT than solely the direct application of electrical current to the body.

One key explanation is that seizure induction affects circulating catecholamines and leads to an increased sympathetic tone [7, 9, 22]. The study by Jones and Knight [22] specifically assessed changes in catecholamines, blood pressure, and heart rate during ECT. They found, in the absence of beta-blockade, there were ischemic ECG changes, an increase in the rate pressure product (peak heart rate times systolic blood pressure) up to 35,000, a 15-fold increase in plasma epinephrine, and a three-fold increase in plasma norepinephrine. A rate pressure product of at least 30,000 is considered a sufficient workload to diagnose coronary disease in cardiac stress testing [23], so these patients effectively undergo a stress test during ECT. Furthermore, increased circulating catecholamines are thought to promote AF through a mechanism of increased magnitude of open L-type calcium channels as well as excessive intracellular calcium [24]. Thus, hyperadrenergic states via an effective stress test can logically lead to rhythm aberrancies.

Another plausible reason for rhythm conversion is the use of antiarrhythmic agents. Some of the patients who experienced cardioversion were also taking agents, including encainide, digoxin, and amiodarone, and one patient failed synchronized cardioversion events peri-procedurally; these factors improve the patient’s propensity to convert [7,8,9].

Notably, the use of many common psychiatric medications has known cardiovascular conduction effects. Although the concern is primarily for QT-prolongation and ventricular events [7, 8], combined with periprocedural anesthetic use, these medications can alter physiologic thresholds, facilitating rhythm changes [7]. Thus, the mechanism of conversion is multi-faceted.

Conversion to sinus rhythm

ECT led to the cardioversion of AF to sinus rhythm in five cases. A majority of the patients who experienced cardioversion to sinus rhythm were elderly with multiple cardiovascular comorbidities, and no cardioversion events led to a stroke in the time these patients were under the care of the respective manuscript’s authors—a duration that was not consistently specified.

The five patients who converted from AF to sinus rhythm are as follows: Harsch [7] described a 75-year-old female with hypertension and AF diagnosed 10 months prior. She spontaneously converted to sinus rhythm 60 s after an ECT session and converted from sinus rhythm to AF on two, separate, later occasions. Ottaway [9] describes a 78-year-old M with coronary artery disease status post 4-vessel coronary artery bypass grafting who developed post-operative stroke, ventricular tachycardia, and atrial fibrillation. He was in AF and anticoagulated with warfarin when he presented for ECT. After his second session, he developed an unspecified, symptomatic tachyarrhythmia associated with pallor and hypotension, so he was started on amiodarone. This event prompted synchronized cardioversion for AF before the next session. It was unsuccessful, so he was started on amiodarone, and ECT proceeded without issue. He subsequently converted from AF to sinus rhythm during the 4th ECT session.

Last, Petrides and Fink [8] describe the conversion of AF to sinus rhythm in three patients: (1) An 89-year-old female with congestive heart failure and AF who was trialed on multiple antidepressant and antipsychotic medications without improvement in her depression. Following her first session of ECT, she converted from AF to sinus rhythm, and during a later session, she reverted to AF. (2) A 76-year-old female with coronary artery disease, congestive heart failure, aortic valve replacement, and hypertension was incidentally found to have new AF pre-ECT, so she was started on warfarin. She alternated between AF and sinus rhythm during her second session and converted to sinus rhythm during her third session. (3) A 68-year-old female with AF and subacute bacterial endocarditis converted to sinus rhythm after her first session of ECT. She subsequently underwent four unremarkable ECT sessions.

Development of atrial fibrillation

There are 17 reported cases of AF following ECT. Three of these cases were in patients who previously converted from AF to sinus rhythm, as described earlier [7, 8]. At least another three cases relate to pre-treatment with the antipsychotics olanzapine, clozapine, and quetiapine, which affect cardiac conduction and repolarization. A 70-year-old male treated with olanzapine had previously tolerated ECT but developed AF after 38 uneventful treatments [12]. Another, a 64-year-old male patient treated with clozapine, developed AF following his second ECT [15]. This case of AF induction was additionally attributed to the use of a charge of 230 mC, 500% greater than his seizure threshold. Narasimhan [13] documented a 70-year-old female taking quetiapine who developed AF with rapid ventricular response and 4-mm diffuse T-wave inversions immediately following bifrontal 50% ECT. This patient converted back to sinus rhythm after administration of esmolol and magnesium sulfate; however, further investigation was required to understand the etiology of her abnormal cardiac findings. Chart review revealed a 200-pack-year history of tobacco use and critical stenosis of his left anterior descending artery. Narasimhan postulated that pre-existing cardiac abnormalities likely contributed to arrhythmia propagation following ECT.

Kuwahara et al. [17] described 3 patients with catatonic schizophrenia who each converted 10–14 min after ECT. A 25-year-old male was pretreated with atropine due to the potential for asystole when ECT is performed on the nucleus of the vagus nerve. This synergistic effect of atropine and ECT stimulation was thought to lead to the development of AF. A 46-year-old and 74-year-old male were both incidentally noted to go into AF 10 min post-ECT, and it was postulated by the authors to be related to “delayed and marked sympathetic excitation...as a consequence of ECT” in patients with catatonic schizophrenia, in particular [17].

Next, AF induction appears to primarily occur during the second and subsequent ECT sessions. Venditti et al. [16] describe an 84-year-old male with hypertension who developed AF after 4 years of bimonthly ECT sessions. Similarly, Loeffler and Capobianco [11] report of a 45-year-old otherwise healthy male patient experienced AF after 25 treatments [11]. Craddock and Gilbert [10] describe a 52-year-old male who experienced AF during his second ECT treatment with a return to normal sinus rhythm spontaneously after three hours. O’Melia [14] describes a 57-year-old male who developed AF during his second ECT session and spontaneously converted back to sinus rhythm after 36 h. Petrides and Fink [8] report on a 64-year-old female with known atrial flutter who developed atrial fibrillation on her sixth session, requiring treatment with digoxin. Last, O’Flanagan and Taylor [19] describe a 54-year-old M who developed documented AF after his 8th session and an irregular pulse following his 9th session, but a confirmatory ECG was not available.

Sattar [18] performed a study of the electrocardiogram (ECG) changes after ECT in 25 patients, noting frequent sinus tachycardia (72% of patients) and one episode of AF induction in an 18-year-old otherwise healthy patient. The patient spontaneously reverted to sinus rhythm after 8 days and without treatment. Olsen et al. [20], describes a 21-year-old male patient who had previously undergone multiple, uneventful ECT sessions but developed AF treated with synchronized cardioversion during one session [20]. Majority of the studied patients undergoing ECT sessions remain in sinus rhythm. Remaining in sinus rhythm (normal rate or tachycardic) is the most common outcome.

Stroke risk and bleeding complications

It is important to consider that many of the included cases predate the widespread use of validated stroke and bleeding risk models and do not reliably disclose the risk factors that physicians currently consider when assessing stroke and bleeding risk, such as hypertension and diabetes. Even though patients may not have been appropriately anticoagulated by today’s standards, only one thromboembolic event was described by Suzuki et al. [25]. A 77-year-old female with hypertension and chronic AF who was non-adherent to anticoagulation suffered an embolic stroke one day after her 43rd ECT session, despite remaining in AF post-ECT. She underwent all 42 ECT sessions without cardioversion or other systemic consequences. This is the only known, published case of a patient with AF developing a stroke post-ECT [25], and it is not related to cardioversion of AF.

Intracranial hemorrhage is a potential—but exceedingly rare—complication of ECT [8, 13, 25]. No intracranial or systemic major bleeding events were described in the case studies included in this review.

ECT and AF: should we anticoagulate?

Stroke risk with cardioversion can be mitigated by appropriate periprocedural anticoagulation, but anticoagulation in ECT still poses concerns for major bleeding events, such as intracranial hemorrhage, despite the relative rarity. Several studies have examined safe anticoagulation use in patients with AF who undergo ECT [8, 26,27,28]. Warfarin and direct oral anticoagulants (DOAC) such as apixaban, rivaroxaban, edoxaban, or dabigatran were analyzed, and they were determined to be safe without major bleeding events post-ECT [8, 26,27,28]. In fact, the patient described by Suzuki et al. [25] may well have avoided stroke had she been adherent to an anticoagulation regimen.

Based on this review, the best course of action is to follow the latest guidelines on atrial fibrillation to guide not only stroke- and bleeding risk assessment but also to assist with anticoagulant choice for each patient based on individual risk factors. Guidelines currently favor DOACs as first-line for stroke prevention in AF, excepting patients with mechanical heart valves, very recent implantation of bioprosthetic material, and valvular AF, highlighting lower rates of bleeding complications associated with DOACs than warfarin [6, 29, 30]. The management of patients who must be on warfarin should follow the latest relevant guidelines (e.g., valvular and atrial fibrillation) to guide target INR selection, and these patients should maintain guideline-direct follow-up intervals with an anticoagulation clinic. Overall, the is no clear contraindication to anticoagulating patients with AF who are at risk for embolic stroke based on validated stroke risk models [6]. In contrast, there is no evidence to support routine anticoagulation before ECT.

Pre-ECT evaluation

Pre-ECT evaluation should follow a similar protocol to those in the most recent guidelines for perioperative cardiovascular evaluation prior to noncardiac surgery [31], bearing in mind that perioperative risks and those associated with ECT are somewhat different. The main similarities include the use of anesthesia and significant periprocedural hemodynamic shifts.

Evaluation should start with a careful history and physical exam performed by a member of the patient’s mental health team. Preemptive consultation of a primary care provider, hospitalist, or cardiologist for risk stratification (akin to preoperative evaluation) is not necessary, would delay care, and would add healthcare expenses to the patient and system. During this initial visit, a patient’s history should be evaluated for prior arrhythmias or arrhythmic symptoms, such as palpitations, chest discomfort, lightheadedness, or weakness. History of cardiovascular implantable electronic devices should be noted, ensuring that a cardiologist is involved in its management. For patients who use clinically validated wearable technology, such as the Apple Watch [32, 33], it is pertinent to ask if an irregular heartbeat has ever been detected. This must be followed by a review of the corresponding rhythm strip(s) to reduce the risk of misdiagnosis from artifact or non-AF irregular rhythms [29, 32]. Of note, current AF guidelines emphasize caution with the use of wearables because many devices are not clinically validated [6, 29]. The 2020 European Society of Cardiology guideline for the diagnosis and management of AF somewhat allows for the use of wearables, stating they can be used to definitively diagnose AF if a “single-lead ECG tracing of ≥ 30 s or 12-lead ECG” demonstrates AF—as reviewed by a “physician with expertise in ECG rhythm interpretation” 29. Non-ECG-based wearables (photoplethysmography) cannot be used to make the diagnosis and would require additional testing. Thus, data from wearable technologies should not be neglected as part of the history. The cardiovascular physical exam should include palpation of bilateral peripheral pulses for rate and rhythm, auscultation for murmurs and gallops, and assessment of blood pressure to improve overall risk stratification and prompt further evaluation if the examination is concerning for AF.

If the history and physical exam is negative for suspicion of arrhythmia, then no further testing would be necessary. This will avoid cascade testing that could, at best, delay necessary treatment for the patient and, at worst, lead to complications from testing. Following the model of preoperative risk stratification, routine 12-lead ECG or echocardiography would not be indicated [31].

If the history and physical exam raise concern for an arrhythmia, a 12-lead ECG should be obtained. If the 12-lead ECG reveals AF or atrial flutter, then initiation of anticoagulation should be based on the most recent atrial fibrillation guidelines [6]. It may be appropriate to consult medicine or cardiology for anticoagulation recommendations and risk stratification, depending on the scope of practice and comfort level of the healthcare team members performing the initial evaluation. If outpatient referral is placed for risk stratification and anticoagulation management prior to the initiation of appropriate anticoagulation, patients would be without appropriate therapy for the duration of the wait for a referral appointment. This may be an extended time, and the patient could have a thromboembolic event in the interim. Thus, using convenient guideline-directed [6] risk stratification tools, such as CHA2DS2-VASc [34], can facilitate appropriate initiation of anticoagulation at the initial evaluation visit where history and physical examination was performed. This process is more streamlined if the patient is hospitalized, since consultants can provide timely recommendations, but a new diagnosis of AF should not prompt hospitalization for an otherwise outpatient procedure.

If the 12-lead ECG does not reveal AF or atrial flutter, it is imperative to weight the risks and benefits of further testing and referrals. Downstream testing may include arrhythmic monitoring, such as a Holter monitor (24–48 h) or event monitoring (up to 30 days). This would delay ECT. Considering that this literature review demonstrated a single stroke in a non-anticoagulated patient who underwent ECT, the risk of thromboembolic events does exist among patients who are not on optimal therapy, but this risk is relatively low. Discussion with an interdisciplinary team, including medicine and cardiology may be helpful to further elucidate risk, especially if ECT is non-urgent. Wearable tools, such as the Apple Watch, may detect an irregular heart rhythm but often do not correlate with AF or atrial flutter [32, 35]. As stated previously, a diagnosis of AF can be made by clinically validated ECG-based devices, but uncertainty of the rhythm should be followed up with additional ECG-based recordings, such as a 12-lead ECG or Holter monitor [29]. Wearables should not be routinely used for diagnosis. Although routine use of these wearables has led to increased diagnosis of AF, the clinical significance remains questionable [32, 35]. Thus, until the guidelines comment further on use of these tools, they have limited role in evaluation beyond assessment of irregular heart rhythm in the patient’s history.

Among patients with AF or atrial flutter who are newly started on anticoagulation and have an urgent indication for ECT (within 30 days), there is insufficient data to either recommend or advise against TEE or cardiac computed tomography. However, the overall risk of thromboembolic events based on this review is low. The risk of further evaluation through structural imaging can increase patient cost, delay care, and lead to incidental findings. The pursuit of these should be personalized to each patient and their cardiac risk factors. Shared decision-making with the patient and interdisciplinary discussion that includes a cardiologist may be considered.

Study limitations

There are several limitations to this literature review. Since the data is derived from a small number of case reports and case studies—six of which are at least thirty years old—it may be difficult to generalize this information to broader, more-modern populations when the ECT dose delivered has decreased. Additionally, home medications and patient comorbidities are not consistently documented. Some psychotropic medications (particularly tricyclic antidepressants) have effects on cardiac conduction, which can affect outcomes. Furthermore, antecedent data on asymptomatic, historical AF cannot always be known, and AF is generally prevalent in the study population. Last, ECT dose also was not consistently reported, and this may affect catecholaminergic surge and propensity for cardioversion. Without documentation of predisposing factors, it remains possible that AF and ECT timing is coincidental.

Future areas of study

This study seeks to provide a starting point for additional research in the field, and several topics require further investigation. To specifically quantify risk, a large, randomized study that assesses patients with known AF for cardioversion to sinus rhythm post-ECT is warranted. Additional research on the catecholaminergic effects of ECT may also elucidate an underlying mechanism for rhythm changes. Last, with improvements in wearable cardiac technology since the publication of the majority of these cases, a screening process may evolve to include ambulatory cardiac event monitoring before ECT. In addition to anticoagulation safety in ECT, the role of pre-screening for AF would be an area for future study since there are no current recommendations on the topic.