From brain to heart: a single-centre study of ictal and post-ictal asystole in epilepsy Authors: Abhishek Gohel, Belinda Philip, Akshaya Raman, Rutul Shah, Tejas Shivarthi, Mahima Sriram, Ashok Pillai, Ramaiah Rajeshkannan, Talal Almashaikhi, Navin Mathew, Siby Gopinath Published: 2026 ABSTRACT AbstractObjectiveThe objective of this study is to explore ictal asystole (IA), a neurocardiogenic phenomenon occurring during epileptic seizures, and its implications in patients with epilepsy. The study aims to describe the clinical presentations, diagnostic challenges and management approaches of IA to contribute to the existing knowledge and improve patient care.MethodsA retrospective analysis was conducted on 15 578 patients undergoing prolonged video-electroencephalogram monitoring over a period of 115 months. Clinical features, electrophysiology, neuroimaging, interventions and outcomes were assessed.ResultsEight patients (0.05%) exhibited IA (six females; mean age: 36.5 years; range: 26–45 years). The mean age at seizure onset was 20.5±11.73 years and mean disease duration from seizure onset was 16±10.66 years. Six patients gave a history of hypomotor features (75%), one exhibited automotor features (12.5%) and one exhibited sensorimotor features (12.5%). Of the 127 ictal events analysed, 47 (37%) progressed to asystole and 80 (63%) had bradycardia without asystole. Bradycardia preceded 95.7% of asystolic events. A significant direct correlation was observed between average duration of IA and average seizure duration (p=0.015). IA was predominantly observed in temporal lobe epilepsy. Falls and myoclonus were observed during asystolic events. All patients survived without requiring resuscitation. Six patients underwent surgical intervention for epilepsy, with varying degrees of success. One patient underwent permanent pacemaker implantation.ConclusionClinicians should consider IA in patients with epilepsy presenting with unexplained syncope or sudden falls, especially when associated with recent changes in seizure semiology. A high index of suspicion is essential to enable timely diagnosis and prevent injury. WHAT IS ALREADY KNOWN ON THIS TOPIC Ictal asystole (IA) is a rare but alarming neurocardiac manifestation of epileptic seizures, characterised by a cardiac pause with an R–R interval ≥3 s occurring during the ictal phase. It has been associated with atonic falls and traumatic injuries and may have possible links with sudden unexpected death in epilepsy. The central autonomic network, particularly temporal-insular pathways, is thought to play an important role in the generation of ictal bradyarrhythmias. Important aspects of IA, including mechanisms of seizure termination, seizure propagation patterns and electroclinical correlates of falls, remain incompletely understood. Ictal asystole (IA) is a rare but alarming neurocardiac manifestation of epileptic seizures, characterised by a cardiac pause with an R–R interval ≥3 s occurring during the ictal phase. It has been associated with atonic falls and traumatic injuries and may have possible links with sudden unexpected death in epilepsy. The central autonomic network, particularly temporal-insular pathways, is thought to play an important role in the generation of ictal bradyarrhythmias. Important aspects of IA, including mechanisms of seizure termination, seizure propagation patterns and electroclinical correlates of falls, remain incompletely understood. WHAT THIS STUDY ADDS This study demonstrates a significant direct correlation between seizure duration and IA duration, supporting the hypothesis that IA is a downstream consequence of sustained seizure propagation through autonomic networks rather than an isolated focal phenomenon. We also observed that atonic falls frequently preceded global electroencephalogram (EEG) suppression, raising the possibility of brainstem-mediated mechanisms explaining ictal falls in addition to cerebral hypoperfusion. This study demonstrates a significant direct correlation between seizure duration and IA duration, supporting the hypothesis that IA is a downstream consequence of sustained seizure propagation through autonomic networks rather than an isolated focal phenomenon. We also observed that atonic falls frequently preceded global electroencephalogram (EEG) suppression, raising the possibility of brainstem-mediated mechanisms explaining ictal falls in addition to cerebral hypoperfusion. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY The study findings emphasise the importance of prolonged video-EEG monitoring with continuous ECG assessment in patients with unexplained falls, atypical syncope or evolving seizure semiology. Future prospective studies evaluating blood pressure recordings and respiratory channels in addition to EEG and cardiac channels would be useful. Genetic testing to uncover channelopathies and baseline autonomic function tests that determine an individual’s susceptibility to vagal influences may also be valuable in identifying other possible aetiologies of IA. The study findings emphasise the importance of prolonged video-EEG monitoring with continuous ECG assessment in patients with unexplained falls, atypical syncope or evolving seizure semiology. Future prospective studies evaluating blood pressure recordings and respiratory channels in addition to EEG and cardiac channels would be useful. Genetic testing to uncover channelopathies and baseline autonomic function tests that determine an individual’s susceptibility to vagal influences may also be valuable in identifying other possible aetiologies of IA. Introduction Ictal arrhythmias refer to cardiac conduction defects that manifest during electrographic seizures. Ictal asystole (IA), a rare arrhythmia during seizures, has garnered attention because of its proposed links to sudden unexpected death in epilepsy (SUDEP).1–4 IA is a rare but alarming neurocardiac manifestation of epileptic seizures, characterised by a cardiac pause with an R–R interval ≥3 s occurring during the ictal phase.5 During seizures with IA, patients often present with loss of consciousness, atonia, sudden falls or traumatic injuries. Since its first description, approximately 200 cases have been reported, likely an underestimate, given the transient and intermittent nature of IA. Reported prevalence varies widely, ranging from 0.1% to 0.4% in patients undergoing video-electroencephalogram (VEEG) monitoring at major epilepsy centres, up to 5%–21% in long-term studies using implantable loop recorders.1 2 6 7 Given that IA is rare and many prior observations remain inconclusive or debatable, further studies are required to elaborate on aspects such as the definitive mechanism, possible relation to SUDEP, the correlation between seizure and asystole duration and the temporal relationship of EEG suppression and atonic falls. In this study, we detail eight IA cases, emphasising their clinical presentations, diagnostic hurdles and treatment strategies. Methodology Study setting This study was conducted in the epilepsy monitoring unit (EMU) of a tertiary healthcare centre. Patients were referred either for diagnosis (distinguishing between epileptic and paroxysmal non-epileptic events, or for further categorisation of the types of seizures and epilepsy syndromes) or for presurgical evaluation. Study definitions Prolonged-VEEG (PVEEG) monitoring was defined as continuous VEEG monitoring for 8 hours or more. Seizure duration was defined as the time interval between the first ictal change or first clinical ictal change (whichever came first) and ictal EEG offset.8 Bradycardia was defined as heart rate <40 bpm and/or >10% decrease of baseline heart rate as per a previous study.1 Asystole was defined as the absence of a ventricular complex for ≥3 s after the onset of seizure.5 The duration of asystole was calculated in seconds from the onset of asystole up to the reemergence of sinus rhythm. Latency of IA was computed as time lapse between EEG ictal onset and ECG asystole onset. Post-IA was defined as asystole which immediately follows ictal EEG offset.2 Atonic seizures, or ‘drop attacks’, were defined as a collapse or fall with little or no warning, following which the patient lies motionless or nearly motionless for a variable time, but often with rapid recovery to their normal state.9 Myoclonus was defined as involuntary, brief shock-­like movements caused by muscular contraction or inhibition.10 Technical considerations The research used digital VEEG 32- or 40-channel setups including Nicvue V.2.9.1 and Natus Database V.9.1.0 and V.7.1.0. Electrode placement conformed to the International 10/20 system with additional temporal placements as needed, displaying a minimum of 18 EEG channels alongside one ECG. Patients initiated the VEEG assessment on antiseizure medications (ASMs) and followed the ASM reduction protocol, unless seizures were more frequent than once per week. The protocol dictated that 30% of all ASMs, with the exception of sedatives (eg, barbiturates, clobazam), be progressively reduced every 24 hours, and continued until sufficient data was captured. All participants were subjected to 3T brain MRI scans as per the HARNESS-MRI protocol.11 Study subjects and data collection We included a total of 15 578 patients for the period of January 2013 to August 2022. Patients with IA were retrospectively identified from a database of PVEEG monitored individuals. Demographic data, details of seizure history and ASMs at admission were obtained from electronic medical records. The seizures were classified as per the clinical semiological classification proposed by Lüders et al.12 Diagnostic evaluations like blood tests, cardiac assessments and MRI scans were standard. Event details, history of atonic falls and myoclonic jerks were noted. EEGs were independently reviewed by three expert epileptologists (SG, AG, RS) who were blinded to the patients’ clinical details and outcomes, and doubtful cases were discussed to achieve concordance. Our team reviewed the EEGs and assessed ictal onset, event frequency and duration, asystole onset characteristics, and any secondary generalisations. ECG data from VEEGs provided information on heart rate and rhythm, asystole duration, timings relative to EEG onset and heart rate variations. We also documented interventions, clinical results and postoperative histopathological outcomes to confirm the underlying diagnosis in ambiguous cases, facilitate correlation with the epileptogenic zone and aid disease prognostication. Serum potassium levels of all patients were tested at the time of evaluation of arrhythmia and other confounding electrolyte abnormalities were excluded. We assessed patients for other confounding factors such as prior cardiac issues, family history of premature death and previous use of known arrhythmogenic drugs (beta blockers, calcium channel blockers, etc). Statistical analysis Statistical analyses were conducted using SPSS V.20. Continuous variables were expressed as mean and SD, with ranges mentioned if relevant. Pearson χ2 test or Fisher’s exact test was used to compare categorical variables. Spearman’s correlation was used to find correlation between seizure and IA duration. A p value less than 0.05 was considered significant. Results During the study period, we analysed 15 578 cases undergoing PVEEG monitoring in our EMU. Eight patients had IA with mean age 36.5 years (26–45 years). Six patients (75%) were female. Seizure frequency varied from 5–6/day to 2–3/year. Average age at seizure onset was 20.5±11.73 years, with a disease duration of 16±10.66 years. Six patients (75%) gave a history of habitual hypomotor seizures, one patient (12.5%) had habitual automotor seizures and one patient (12.5%) had sensorimotor seizures. Secondary generalisations were noted in the past in five patients (62.5%) in varying frequencies. Seven patients (87.5%) reported a history of falls. Clinical features are summarised in table 1. Table 1Patient characteristicsCase No.Age/sexHandednessDisease duration (years)At presentationPostoperative seizure statusLuder’s classification of habitual seizuresSeizure frequencyASMsBrain MRI findings1Mid-40s FemaleRight31Hypomotor4–5/monthPHT, ZNSRight parieto-temporal-occipital encephalomalaciaEngel 12Early-40s MaleRight15Hypomotor2–3/yearCLB, LEVNormalEngel 23Mid-40sMaleRight2Hypomotor1/monthVPA, CLBNormalNA4Mid-30s FemaleRight24Hypomotor1–3/monthVPA, CLBLeft sided hippocampal sclerosisEngel 15Mid-30s FemaleRight6Automotor1 cluster/weekLCM, LEVNormalNA6Mid-30s FemaleRight21Hypomotor1/weekLEV, LCM, CBZFCD at deep sulcus posterior to left angular gyrusEngel 27Early-30s FemaleRight24Sensorimotor5–6/dayLEV, LCM,CLBFCD at post-central gyrusEngel 38Mid-20s FemaleRight5Hypomotor4–5/month, clusters up to 12/dayOXC, LCMNormalEngel 1ASM, antiseizure medication; CBZ, carbamazepine; CLB, clobazam; FCD, focal cortical dysplasia; LCM, lacosamide; LEV, levetiracetam; NA, not available; OXC, oxcarbazepine; PHT, phenytoin; VPA, valproate; ZNS, zonisamide. Patient characteristics ASM, antiseizure medication; CBZ, carbamazepine; CLB, clobazam; FCD, focal cortical dysplasia; LCM, lacosamide; LEV, levetiracetam; NA, not available; OXC, oxcarbazepine; PHT, phenytoin; VPA, valproate; ZNS, zonisamide. All were right-handed, on at least two ASMs and had failed three prior medication trials. None of our patients routinely used any known pro-arrhythmogenic medications. No patient had prior cardiac issues or family history of premature death. All patients had a normal interictal ECG, and 2D ECHO findings were normal in 5/8 patients (67.5%). ECHO was not performed in the other three patients. Online supplemental file 1 details all the case vignettes of our cohort. Clinical semiology ASM reduction protocol was required in 7/8 patients (87.5%) as case 7 had frequent seizures despite ASM use. On VEEG recording, seven patients (82.5%) displayed hypomotor seizures, with case 5 exhibiting new hypomotor seizure semiology. One patient (12.5%) exhibited sensorimotor seizures. In the patients with hypomotor seizure presentation, 2/7 (28.6%) had seizures evolving into automotor presentation, 1/7 (14.3%) showed non-versive left-sided head and eye deviation and 1/7 (14.3%) showed version to the right side. Atonic falls were observed in three patients (37.5%). Electrophysiology This study identified varied ictal onset locations, including the left temporal lobe in four patients (50%), right temporal lobe in three patients (37.5%) and left central region in one patient (12.5%). Table 2 describes the features of analysed seizures and asystole as per VEEG. A total of 127 ictal events were studied among the eight patients who had IA, of which 47 were asystolic (37%), while 80 (63%) were bradycardia without progressing to asystole (figure 1). The ratio of ictal to asystolic events ranged from 1:1 to 1:13 in our series. Notably, one patient (case 7) exhibited 31 of the 47 events of IA recorded by VEEG. Additional details are included in online supplemental file 3. Electrophysiological correlates. Table 2Features of analysed seizures and asystoleCase No.Location of ictal onsetPre-ictal HRChange in HRRatio of asystolic events to total events (as recorded by VEEG)Luder’s classification of seizure at ictal onsetSequence in seizure semiologyTotal duration of ictal EEG(s)Duration of asystole(s)1*RT66Bradycardia5:6HypomotorHypomotor→Atonic fall56142RPT78Bradycardia4:4HypomotorHypomotor→Non-versive LHD→Irregular bilateral convulsive movements82113*LT60Bradycardia1:3HypomotorEpigastric aura→Hypomotor4864LT54Bradycardia1:2HypomotorHypomotor→Automotor→Fall77285LT80Bradycardia4:4HypomotorHypomotor→Bilateral convulsive movements49286LFT96Tachycardia2:2(both events were postictal asystole)HypomotorHypomotor→Version to right→Post-ictal stupor2751457LCP60Bradycardia31:93SensorimotorSensorimotor left→Non-versive LHD54138RT72Bradycardia1:13HypomotorHypomotor→Automotor (no fall)→Axial tonic posturing444*Events where clinical onset of ictus occurred before ictal onset.HR, heart rate; IA, ictal asystole; LCP, left centro-parietal region; LFT, left fronto-teporal region; LHD, left head deviation; LT, left temporal region; RPT, right posterior temporal region; RT, right temporal region; VEEG, video-EEG. Features of analysed seizures and asystole Events where clinical onset of ictus occurred before ictal onset. HR, heart rate; IA, ictal asystole; LCP, left centro-parietal region; LFT, left fronto-teporal region; LHD, left head deviation; LT, left temporal region; RPT, right posterior temporal region; RT, right temporal region; VEEG, video-EEG. Bradycardia preceded 45 asystolic events (95.74%), while two events in case 6 had preceding tachycardia. The mean duration of IA was 12.13±10.02 s (4–287 s). Mean duration of the seizure in asystolic events was 65.96±20.92 s (49–493 s) and the mean latency to asystole onset was 25.94±9.91 s (7–202 s) after EEG onset. In comparison, the mean duration of seizure in non-asystolic events was 66±26.45 s (6–161 s). Correlation analysis between average durations of IA and seizure demonstrated a significant direct relationship (p=0.015) (figure 2). Scatter plot between average duration of ictal asystole (IA) and average seizure duration. In case 6, there were two asystolic events in the post-ictal period lasting 287 s and 145 s respectively. Tachycardia in this patient began at ictal onset and continued post-resolution of the ictal activity, eventually culminating in post-IA. Careful review of recordings did not demonstrate apnea preceding the cardiac arrest. Instead, the patient exhibited abnormal agonal gasping respirations immediately before the onset of asystole. For all but two patients, we recorded less than 10 total events per patient, an average of 2.67 asystolic events per recording and an average recording duration of 5±2.55 days (2–9 days). In case 7, there were 93 total events recorded over 8 days, with 31 of those being asystolic events with an average duration of 8.5±0.60 s. Epileptiform activity on EEG was seen after the resolution of IA in a combined 11 events in cases 2, 4 and 7. Nine out of 47 asystolic events were followed by global suppression of EEG, with an average time of 49.67±7.43 s (37–58 s) from ictal EEG onset to global suppression. Diffuse EEG slowing preceded global suppression in all nine cases where it was seen. Five events of global suppression were preceded by a clinically recognisable atonic fall. The mean time to diffuse slowing from electrical onset was 26.17±18.63 s. In all patients, sinus rhythm returned spontaneously, and there was no need for cardiac resuscitation. We found one instance of multifocal myoclonus following a generalised seizure, which coincided with the resumption of electrical activity as measured by scalp EEG. Neuroimaging In our cohort of patients, four (50%) had no brain MRI abnormalities. Of the four MRI-positive cases, one exhibited right parieto-temporal-occipital encephalomalacia (online supplemental file 4), one showed left hippocampal sclerosis and two had focal cortical dysplasia (FCD). No cases involved frontal lobe, cingulate, insula or orbitofrontal regions. Intervention and outcome In our cohort, six patients (75%) received surgical treatment for epilepsy. Case 1 underwent right posterior quadrantic disconnection and remains Engel 1 at 11-year follow-up with no subsequent falls. Case 2 underwent right posterior inferior temporal focal cortical resection and is Engel 2 without generalisations or falls at 8-year follow-up. Case 4 had a left anterior temporal lobectomy and amygdalohippocampectomy and remains Engel 1 at 5-year follow-up. Case 6 received a left lateral occipital FCD resection and stands at Engel 2 at 9-month follow-up. Case 7 underwent left parietal FCD resection, however asystolic events persisted, and she underwent a permanent pacemaker (PPM) implantation. She is Engel 3 after 26 months. Case 8 underwent right anterior temporal lobectomy and amygdalohippocampectomy and is Engel 1 postsurgery at 6-month follow-up. The other two patients chose to optimise their ASMs and pursue medical management. Discussion To the best of our knowledge, this is one of the larger single-centre case series on IA. The largest number of IA cases previously reported in a single centre was documented by Schuele et al, Bestawros et al and Khalil et al, with each study reporting 10 cases.1 13 14 Clinical features Our cohort did not demonstrate a clear hemispheric lateralisation, with asystolic events occurring in seizures originating from either hemisphere. Earlier hypotheses suggest a predominance of left temporal seizures in ictal bradyarrhythmias, attributed to greater parasympathetic cardiac control mediated by the left insular cortex.15 However, subsequent clinical studies have yielded inconsistent findings.6 Several later studies have documented IA occurring with both left and right-sided epileptogenic foci. These observations suggest that IA likely reflects complex interactions within the central autonomic network rather than a strictly hemispheric effect.2 Among our cohort, case 6 had a period of asystole lasting 287 s and remarkably survived. Such cases have been rarely reported in the literature, with classic risk factors for this presentation primarily attributed to medically refractory epilepsy secondary to temporal lobe seizures. A previous study by Beach et al, reports a patient with an IA duration of 50 s, for whom PPM implantation was pursued.16 Other case reports and small studies have concluded that PPM implantation in patients who experience IA >6 s decreases the risk of future syncopal episodes and frequent falls, improving mortality and morbidity.17 In case 5, multifocal myoclonic jerks commenced immediately on return of cardiac rhythm following IA, which we describe as post-anoxic, recovery-phase myoclonus, analogous to reperfusion arrhythmias, where cardiac rhythm disturbances occur time-locked to coronary reperfusion following thrombolysis in myocardial infarction.18 We hypothesise an acute post-hypoxic cerebral hyperexcitability, where reperfusion after transient anoxia initiated a transient window of oxidative stress and altered neuronal excitability, representing a neurological correlate of reperfusion at the EEG–cardiac rhythm interface.19 Epidemiology Sinus tachycardia is the most common arrhythmia in epileptic seizures.2 Ictal bradycardia occurs in fewer than 6% of seizures.20 IA prevalence varies from under 0.4% in VEEG monitoring units to 16% in long-term ECG monitoring by implantable loop recorder, suggesting a potentially higher prevalence than estimated by VEEG.3 6 Although more than 15 000 patients were screened in our study, the number of patients with IA was eight, which is lesser than that observed in literature. Only a subset of the screened patients in our study who required surgery underwent PVEEG monitoring for more than 24 hours, hence the true prevalence may be underestimated. In our series, the ratio of ictal events to asystolic events varied among patients, ranging from 1:1 to 13:1. This highlights the importance of capturing the seizures of all different semiologies in a patient, particularly in the context of a recent change in the clinical presentation. It also highlights the fact that not every event in a given patient needs to be an asystolic event. Risk factors In our study, IA showed female preponderance, correlating well with previous studies.3 Seven of our patients had temporal lobe epilepsy (mesial and neocortical), in line with existing literature describing temporal lobe preponderance in IA and that these patients are more prone to autonomic dysfunction.4 15 21 Additionally, none of our patients had any medical comorbidities, including cardiac issues. Pathophysiology The pathophysiology of IA remains unclear. Stimulation of insula, cingulate, amygdala, limbic system and orbitofrontal cortex has been known to produce asystole.16 22 23 Furthermore, temporal autonomic networks have been implicated in heart rate variability during seizures.24 All the asystolic events in our study started as bradycardic events, except in one patient where it began as tachycardia during the generalised phase of ictus and progressed to post-IA. The duration of asystole in this patient was extraordinarily long and occurred early in our unit’s experience with peri-ictal cardiac arrhythmias, before dedicated alarm systems had been implemented. Consequently, the significance and urgency of the ECG changes were not immediately appreciated in real time. Following the event, automated ECG alarms for heart rates below 40 bpm and above 120 bpm were introduced to facilitate prompt recognition and intervention. Post-IA is believed to be the effect of significant cardiorespiratory inhibition and is typically preceded by postictal generalised EEG suppression and associated with convulsive seizures.2 25 These characteristic patterns were not documented in our patient. This raises the possibility of an abnormal brainstem-mediated cardiorespiratory reflex contributing to the event, in addition to cerebral hypoperfusion. This contrasts with the sequence commonly reported in association with SUDEP, where seizures culminate in either immediate death or transient partially restored cardiorespiratory function followed by terminal apnea and consequent cardiac arrest.26 In our study, we found that IA occurred predominantly in patients with temporal lobe epilepsy, correlating well with the findings of Khalil et al.14 While these asystolic events are not commonly seen in patients with epilepsy, they have been proposed to confer a risk of premature death, with seizure-associated cardiac rate and rhythm abnormalities being implicated in SUDEP.2 Importantly, cerebral anoxic ischaemia, an expected consequence of IA, has been suggested to terminate ongoing ictal activity, forming the basis of the hypoperfusion hypothesis, and argues against IA as a primary driver of SUDEP.27 Interestingly, our cohort described significant direct correlation between the average duration of asystole and average seizure duration. A longer duration of asystole correlated with prolonged seizure duration, favouring a neurogenic mechanism. This supports the hypothesis that sustained ictal activity engages the central autonomic network, resulting in excessive vagal output to the cardiac conduction system and consequent sinus arrest.28 Taken together, our findings suggest that IA is likely a downstream effect of prolonged seizure activity although the potential contribution of cerebral hypoperfusion in eventual seizure termination cannot be excluded. The fact that all our cases survived their events could further support the role of cerebral anoxia in premature seizure termination, thereby lending additional weight to the hypoperfusion hypothesis. An intriguing finding was that the time from asystole onset to atonic fall was shorter than that to global suppression. This may suggest that the loss of cerebral blood flow might not be the predominant mechanism for atonic fall, as falls occurred earlier than EEG suppression. We also analysed the ictal EEG pattern at the onset of asystole. In all patients, seizure activity had already propagated beyond a unilateral focal onset, involving either bilateral temporal regions or diffuse/generalised spread at the time of asystole. This supports prior observations suggesting that IA is more likely related to seizure propagation through broader cortical-subcortical autonomic networks rather than activation of a single isolated focus.29 The MORTEMUS study proposed SUDEP as a possible end result of a centrally mediated neurovegetative breakdown, leading to either an immediate death or transient partially restored cardiorespiratory function leading to a terminal apnea followed by cardiac arrest. In this study, one patient with IA experienced a near-SUDEP event, wherein early CPR might have prevented the fatality.26 Similarly, in our current study, no deaths occurred during monitoring, aligning with previous research suggesting that IA is self-limiting.27 The routine use of cardiac monitoring, skilled nursing care and early primary measures in our EMU may have contributed to the seemingly benign nature of the monitored IA. Epileptic heart Patients with chronic epilepsy may have undetected arrhythmia warranting a regular cardiac check-up.30 Chronic epilepsy can lead to recurrent catecholaminergic overstimulation, promoting myocardial ischaemic damage and fibrosis, predisposing to myocardial dysfunction and cardiac arrhythmia.31 None of our patients had any cardiac abnormalities after extensive evaluation with ECG, ECHO, treadmill test and Holter investigations. Similar findings were reported in a previous study by Schuele et al.1 Postulated mechanisms Several factors are hypothesised to contribute to IA: The lockstep mechanism describes the intermittent synchronisation of cardiac sympathetic and vagal neural discharges as a factor in cardiac arrhythmias.32 The lockstep mechanism describes the intermittent synchronisation of cardiac sympathetic and vagal neural discharges as a factor in cardiac arrhythmias.32 The inhibitory effect of the cardiorespiratory reflex on the heart may reflect limbic system activation in emotional stress, as seen in vasovagal asystole. As noted by Schuele et al, the similarities in heart rate characteristics of IA and vasovagal asystole raise the possibility of medullary centres mediating increased vagal tone in IA as well.33 The inhibitory effect of the cardiorespiratory reflex on the heart may reflect limbic system activation in emotional stress, as seen in vasovagal asystole. As noted by Schuele et al, the similarities in heart rate characteristics of IA and vasovagal asystole raise the possibility of medullary centres mediating increased vagal tone in IA as well.33 Genetic channelopathies may predispose patients to cardiac arrhythmia, especially if the affected ion channel is expressed in the brain, leading to coexisting epilepsy.34 Genetic channelopathies may predispose patients to cardiac arrhythmia, especially if the affected ion channel is expressed in the brain, leading to coexisting epilepsy.34 Drug-induced factors, including agents like beta blockers, verapamil, digoxin and ASMs such as carbamazepine, can influence the QT interval, increasing the risk of IA.35 Drug-induced factors, including agents like beta blockers, verapamil, digoxin and ASMs such as carbamazepine, can influence the QT interval, increasing the risk of IA.35 Challenges in diagnosis Diagnosis of IA poses challenges as they are often misinterpreted as a syncope.13 Loss of tone with diffuse slowing is known to be seen in syncope, and similar findings are documented in IA, causing confusion.36 A possible explanation for these clinical manifestations of IA was proposed by Ghearing et al, where it was suggested that cerebral hypoperfusion secondary to asystole was likely the cause, as opposed to stimulation of cortical and sub-cortical regions by a seizure.37 IA may be a relatively benign phenomenon compared with cardiac asystole, as all the patients in this study have survived, including those with prolonged and frequent events, without apparent consequences. It is possible that a negative feedback loop exists, with the exact mechanism remaining unclear. Even in cases with prolonged (case 6) and frequent events (case 7), the outcomes were benign. However, IA is associated with significant morbidity, including sudden falls and potential head injuries.38 Consequently, cardiac pacing is recommended for IA lasting 6 s or more, as they have been associated with clinical falls.39 Hypomotor events often go unnoticed, only being recognised when they present with anoxic manifestations, as in case 5, who was apparently seizure-free for 3 years, but truly achieved seizure freedom after resective surgery. Clinicians should consider IA in the differential diagnosis of unexplained syncope or loss of consciousness in patients with epilepsy, particularly when accompanied by falls or sudden change in seizure semiology.40 As the field of IA research continues to evolve, future research should aim to deepen our understanding of the phenomenon by evaluating blood pressure recordings and respiratory channels in addition to EEG and cardiac channels. Genetic testing to uncover channelopathies and baseline autonomic function tests that determine an individual’s susceptibility to vagal influences may be employed to identify other possible aetiologies of IA. Supplementary material