The Heart-Brain Axis: Key Concepts in Neurocardiology

Abstract

The heart-brain axis involves complex interactions between the cardiovascular and nervous systems via a network of cortical and subcortical structures working with the autonomic nervous system and intracardiac nervous system. Heart-brain interactions may be divided into 2 broad categories: cardiac effects of neurological disease and neurological effects of cardiac disease. The pathogenesis of neurogenic cardiac effects is thought to involve a neurogenic cascade where sudden shifts in autonomic balance lead to an exaggerated catecholamine release. This can occur in acute neurological conditions such as ischemic stroke, intracranial hemorrhage, and epilepsy. Cardiovascular complications include the stroke-heart syndrome, neurogenic pulmonary edema and cardiomyopathy, Takotsubo syndrome, arrhythmias, and even sudden cardiac death. Certain areas of the brain, such as the insular cortex, play key roles in cardiac autonomic regulation, and disorders affecting these areas have greater effects on the heart. On the other hand, cardiac conditions can also adversely impact the neurological system. Atrial fibrillation and left ventricular thrombus can cause cardioembolic strokes, whereas heart failure and severe aortic stenosis have been linked to the development of cognitive impairment. This review aims to provide a broad overview of key topics in neurocardiology as well as delve into the evidence and pathophysiology behind these conditions.

1. Introduction

The existence of a close relationship between the heart and the brain has been known for many years. Physiologist Walter Cannon first described numerous instances of death from fright seen in anthropology literature, which he termed “Voodoo Death” in his paper published in 1942.^[1] Cannon hypothesized that these extraordinary deaths were caused by activation of the sympatho-adrenal system. Later in 1971, George Engel, an internist and psychiatrist, classified similar cases of sudden death into 8 categories: (1) the impact of the collapse or death of a close person, (2) during acute grief, (3) threat of loss of a close person, (4) during mourning or on an anniversary, (5) on loss of status or self-esteem, (6) personal danger or threat of injury, (7) after the danger is over, and (8) reunion, triumph, or a happy ending.^[2] He observed that these events all involved overwhelming despair, excitement, or both, and postulated that these emotions provoked neurological responses conducive for lethal cardiac events. In more recent times, ventricular arrhythmias detected on implantable cardioverter-defibrillators increased by more than 2-fold among patients with implantable cardioverter-defibrillators following the World Trade Center attack in 2001.^[3] Similarly, a Japanese study reported a doubling of the incidence of sudden cardiac and unexpected death in the initial 4 weeks after the 2011 Japan earthquake and tsunami. This effect was more pronounced in residents who lived in the tsunami-stricken area, and the incidences returned to prior levels following the immediate aftermath.^[4] These observations over decades shed light on the potential effects of acute stress on inducing arrhythmogenesis and sudden death.

It is now known that a network of cortical and subcortical structures, together with the autonomic nervous system, interact to form the heart-brain axis.^[5–7] These interactions may be divided into 2 broad categories: cardiac effects of neurological disease and neurological effects of cardiac disease [Figure 1]. Certain areas of the brain such as the insular cortex play key roles in autonomic regulation of the heart, and pathology in these areas can cause significant cardiac consequences.^[8–10] This review article aims to provide a broad overview of key topics in neurocardiology, as well as delve into the evidence and pathophysiology behind these conditions.

Figure 1.

Heart-brain interactions. Images created with BioRender.com. AF: Atrial fibrillation; ECG: Electrocardiographic; ICH: Intracranial hemorrhage; LV: Left ventricular.

2. Neural regulation of cardiovascular function

Neural modulation of cardiovascular function involves a complex network of cortical and subcortical structures. This network comprises cortical regions such as the medial prefrontal cortex and insular cortex; subcortical structures such as the amygdala and hippocampus; and brain stem areas, which include the hypothalamus, bed nucleus of the stria terminalis, periaqueductal gray, parabrachial region, and rostral ventrolateral medulla. Ascending afferent cardiac neurons carry information to the higher cortical areas, which then process this information and modulate efferent autonomic outflow to regulate cardiovascular function [Figure 2].^[5,6]

Figure 2.

Overview of structures involved in the neural regulation of cardiovascular function. Images created with BioRender.com.

2.1. Insular cortex

Within this central nervous system network, the insular cortex has a crucial role demonstrated in animal and human studies. The insular cortex makes up 2% of the total cortical surface area and is located at the lateral aspect of the forebrain, at the base of the sylvian fissure. It is divided into posterior and anterior lobules by the central sulcus of the insula.^[11,12] Cardiac chronotropic sites were demonstrated in the rat insular cortex, where stimulation of the rostral posterior insula induced tachycardia, and stimulation of the caudal posterior insula resulted in bradycardia.^[13] The role of the insular cortex in autonomic regulation was further established in a study that found significantly elevated catecholamine levels following induced insular infarction in cats compared to pre-infarction levels. Such elevations were not observed in infarctions which did not involve the insula. A possible explanation is that the attenuation of inhibitory feedback from the insula on central cardiovascular regulating centers following infarction resulted in decreased sympatho-adrenergic activity.^[14]

There appears to be hemispheric lateralization of autonomic activity in primates. Baroreceptor challenge using phenylephrine hydrochloride and sodium nitroprusside revealed a larger number of baroreceptor-related neurons within the right insula compared to the left in primates.^[15] Similarly in humans, the right insular cortex is involved in sympathetic nervous activity, whereas the left insular cortex mediates the vagal nervous system and parasympathetic tone.^[16,17] In provocation studies, intraoperative stimulation of the left insular cortex in humans frequently produced bradycardia and depressor responses whereas stimulation of the right insular cortex resulted in tachycardia and pressor effects.^[18] Among patients with middle cerebral artery stroke, pathological activation of the sympathetic nervous system was greatest in those with right hemispheric stroke involving the insular cortex.^[8] Right-sided insular stroke has also been associated with lower heart rate variability.^[19] These studies reinforce the important role of the insular cortex in regulating autonomic activity.

2.2. Subcortical areas and brain stem

A study which used functional magnetic resonance imaging to measure parallel changes in heart rate and fluctuations in local blood flow showed that the amygdala was important in controlling heart rate in negative affect while the hypothalamus had a key role in regulating heart rate in positive affect.^[20]

The amygdala is part of the limbic system and forms part of the mesial temporal lobe. It is found anterior to the hippocampus and indents the tip of the temporal horn. While the amygdala is best known for its role in processing aversive information, it is involved in emotional responses to various sensory stimuli.^[21] The amygdala gives off projections to the prefrontal and orbitofrontal cortex, and it forms reciprocal connections with the hypothalamus via the fornix and stria terminalis. In doing so, it modulates the autonomic effects of emotional stimuli on the heart.^[22–24]

The hypothalamus is involved in transmitting autonomic information to the brain stem via the hypothalamic nuclei. These nuclei include the nucleus tractus solitarius, periaqueductal gray, parabrachial region, rostral ventrolateral medulla, and dorsal motor nucleus of the vagus.^[25] Different areas of the hypothalamus are involved in modulating distinct autonomic responses. For example, parasympathetic responses appear to be specifically mediated by the anterior hypothalamus, and stimulation of this area in animals often result in bradycardia,^[26–28] whereas sympathetic responses such as tachycardia are largely regulated by the lateral hypothalamus.^[27] Cardiac effects of anterior and lateral hypothalamic stimulation are abolished with sympathectomy and vagotomy, respectively, further demonstrating their distinct roles.^[27,28] Apart from changes in heart rate, the hypothalamus may also be involved in the development of arrhythmias. Atrioventricular dissociation, frequent premature systoles, and ischemia-like electrocardiogram (ECG) changes were observed in lateral hypothalamic stimulation, whereas nodal rhythms, aberrant ventricular conduction, and fusion beats were seen in posterolateral hypothalamic stimulation.^[26]

2.3. Cardiac autonomic nervous system

In 1914, Levy^[29] described the obliteration of chloroform anesthesia-induced ventricular tachyarrhythmias in animals by cardiac sympathetic denervation. He also found that sympathetic activation produced similar ventricular tachyarrhythmias, indicating direct neurological mediation of these abnormal rhythms. Neural control of the heart, consistent with Levy’s observations, is under the influence of the autonomic nervous system, which comprises the sympathetic and parasympathetic nervous systems. It is mediated by a series of reflex control networks, which carry information from higher centers such as the insular cortex, anterior cingulate cortex, medial prefrontal cortex, and amygdala to the brain stem. From the brain stem, pre-sympathetic efferent neurons project to the sympathetic branch and then to the pre-ganglionic neurons in the intermediolateral column of the spinal cord. These efferent connections then synapse to the post-ganglionic sympathetic neurons in the intrathoracic ganglia and, finally, to the heart via projections to the intrinsic cardiac ganglia.^[7,30,31] Parasympathetic innervation of the heart starts with pre-ganglionic neurons in the nucleus ambiguous of the brain stem, travels along the vagus nerve, and then projects to post-ganglionic neurons in the intracardiac nervous system ganglia.^[5,7] Both sympathetic and parasympathetic efferent neurons to the heart assemble within these atrial and ventricular ganglia, where they exert control on cardiac electrical and mechanical activity.^[7]

The intracardiac nervous system is sometimes referred to as the “brain within the heart” or the “little brain” of the heart.^[32] It consists of a series of interconnected ganglionated plexi distributed throughout the atrial walls and around the atrioventricular border. This network is made up of sensory (afferent), interconnecting, and cardio-motor (efferent) neurons, which modulate central nervous system control of the heart.^[7,30,33] A study of rats’ hearts using retrogradely transported tracers injected into the animals’ left or right ventricles demonstrated that different groups of intracardiac ganglion cells project to distinct regions of the heart.^[34] Specific electrical or chemical stimulation of selective loci in the intracardiac ganglia has also been shown to give rise to bradycardia, tachycardia, or a biphasic response of bradycardia followed by tachycardia. Bradycardic effects were abolished by atropine and tachycardiac effects by beta-blockade, further supporting the autonomic origin of these changes in heart rate.^[35] Although individual ganglionic plexi are involved in specific autonomic functions, there is evidence of interconnections and interplay among the various plexi to modulate sinoatrial nodal and atrioventricular nodal function.^[36] The autonomic nervous system also exerts influence on the heart via neurotransmitters. Sympathetic neurons release norepinephrine, which, together with epinephrine released by the adrenal glands, activates adrenergic receptors in the heart. Similarly, parasympathetic neurons release acetylcholine, which activates muscarinic acetylcholine receptors.^[31] The heart and brain are therefore linked via a hierarchical system of autonomic control, and this results in both physiological and pathological consequences on either organ system when the other is affected.

3. Cardiac effects of neurological disease

3.1. Stroke-heart syndrome

Acute ischemic stroke has been associated with a multitude of cardiovascular complications, and cardiovascular disease is the most common cause of death in stroke survivors.^[37] A retrospective cohort study of 365,383 patients with ischemic stroke showed that 11.1% subsequently developed acute coronary syndrome, 8.8% atrial arrhythmias, 6.4% heart failure, 1.2% severe ventricular arrhythmias, and 0.1% Takotsubo syndrome within 4 weeks of stroke.^[38] Although stroke and heart disease share similar risk factors, post-stroke cardiac events occur even in patients without known comorbid cardiac disease.^[39] Cardiac events also appear to peak in the immediate post-stroke period,^[39–42] and greater stroke severity is associated with increased risk or degree of post-stroke cardiac complications.^[40,43] Studies using experimental animal models have also demonstrated cardiac complications following induced stroke.^[44–47] Overall, the evidence suggests that acute stroke itself is involved in the pathogenesis of cardiac disorders, and this phenomenon is termed the “stroke-heart syndrome”.

Stroke-heart syndrome is thought to involve a neurogenic cascade, where acute shifts in autonomic balance lead to increased sympathetic tone and exaggerated catecholamine release [Figure 3].^{[14,44,48–50]} Significant alterations in heart rate variability have also been identified in patients with acute stroke.^[51] These autonomic effects result in hyperactivity of calcium channels and increased sarcomeric contraction. On histology, this is seen as contraction band necrosis or coagulative myocytolysis, which is associated with adrenergic stress including in Takotsubo cardiomyopathy.^[52–54] Sympathetic surge may also cause coronary vasospasm and tachyarrhythmias, both of which can result in type 2 myocardial infarction from demand-supply mismatch. Conversely, altered parasympathetic activity may increase vagal tone and trigger bradyarrhythmias, which eventually lead to demand ischemia.^[48] These mechanisms likely underlie the high incidence of myocardial injury seen in patients with stroke.^[55] Some brain structures have a prominent role in autonomic control of the heart, and infarction in these areas is especially implicated in the stroke-heart syndrome. For example, voxel-based lesion-symptom mapping in 299 patients with magnetic resonance imaging-confirmed acute anterior circulation stroke revealed that infarctions involving the anterior insular cortex of the right hemisphere were significantly associated with relative temporal changes of cardiac troponins. This was not seen for other lesions within the anterior circulation in the same study.^[56] Another study showed that acute infarctions in the right insular, right frontal, and right parietal cortex; right amygdala; basal ganglia; and thalamus were associated with post-stroke cardiac arrhythmias.^[57]

Figure 3.

Pathophysiology and classification of the stroke-heart syndrome. Images created with BioRender.com. ECG: Electrocardiographic; LV: Left ventricular.

Systemic and local inflammatory responses may also be implicated in the stroke-heart syndrome. A study in mice demonstrated increased granulocytes, apoptotic cells, and upregulation of pro-inflammatory cytokines within myocardial tissue following transient middle cerebral artery occlusion. These animals had 4-fold elevated troponin levels as well as systolic left ventricular dysfunction with impaired global longitudinal strain and reduced cardiac output compared to controls.^[45] In another experimental mouse model, stroke significantly increased macrophage infiltration into the heart, macrophage-associated inflammatory cytokine levels, as well as induced cardiac fibrosis and hypertrophy. Splenectomy was found to attenuate these effects, suggesting that the spleen may be involved in this immune response.^[58] The role of inflammation in stroke-mediated cardiac injury in humans is less known, although changes in splenic volume and elevated cytokine levels have been described in post-stroke patients.^[59]

Stroke-heart syndrome is classified into 5 main groups: (1) ischemic and non-ischemic asymptomatic acute myocardial injury presenting with elevated cardiac enzymes; (2) post-stroke acute myocardial infarction; (3) left ventricular dysfunction, heart failure, and post-stroke Takotsubo syndrome; (4) ECG changes and cardiac arrhythmias including post-stroke atrial fibrillation (AF); and (5) post-stroke neurogenic sudden cardiac death.^[60]

3.1.1. Myocardial injury and infarction

Stroke-related myocardial injury has been previously reported in animal studies.^[44,45] These experimental models showed elevated cardiac enzyme levels accompanied by left ventricular dysfunction with impaired global longitudinal strain, reduced left ventricular fractional shortening, and lower cardiac output.^[44,45] In humans, a raised cardiac troponin level is a common finding in ischemic stroke^[55] and is likely to reflect similar ischemic or non-ischemic (non-coronary) myocardial injury.^[48,61] Ischemic myocardial injury includes type 1 myocardial infarction secondary to coronary plaque rupture, and type 2 myocardial infarction from a demand-supply mismatch in myocardial oxygenation. Demand-supply mismatch may occur in associated AF, hypertensive emergency, and increased sympathetic outflow in patients with pre-existing coronary artery disease. Non-ischemic myocardial injury includes that of a neurogenic etiology in the stroke-heart syndrome, such as acute heart failure and stress cardiomyopathy, as well as other concomitant non-neurogenic causes of elevated cardiac enzymes.^[48] Overall, myocardial injury without infarction occurs more frequently.^[61–63] A study of 2,123 patients with acute ischemic stroke showed that 13.7% of these patients had elevated cardiac troponin levels >50 ng/L. Although troponin levels were elevated to a similar extent when compared with matched patients with non-ST elevation acute coronary syndrome, patients with acute ischemic stroke were less likely to have coronary culprit lesions or obstructive coronary artery disease on invasive coronary angiography.^[62] Furthermore, in the patients with stroke and culprit lesions on coronary angiography, the majority had grade 3 flow beyond the culprit artery, as measured by the Thrombolysis In Myocardial Infarction grading system. There are no formal guidelines on the management of elevated cardiac enzymes in the setting of stroke, although most of these patients are treated conservatively during the acute phase.^[62] This is especially true for those with large infarctions or a high risk of hemorrhagic conversion.

Beyond the immediate post-stroke period, an invasive approach may be considered, and strategies such as using drug-eluting balloons instead of stents, or a shorter duration of dual antiplatelet therapy, may be employed in patients with high bleeding risk.^[64] Even when managed conservatively, clinicians should be mindful that troponin elevation in acute ischemic stroke is not benign, and even moderately elevated cardiac enzyme levels have been associated with poor outcomes.^[65,66] A prospective observational study of 562 patients with first-ever stroke showed that those with high-sensitivity cardiac troponin elevations above the upper reference limit were more likely to develop a subsequent major vascular event, defined as a composite of recurrent stroke, myocardial infarction or all-cause death (27.3% vs. 10.2%, adjusted hazard ratio: 2.0, 95% confidence interval: 1.3–3.3).^[67] Another study showed that rising cardiac troponin levels within 48 h of acute ischemic stroke were associated with an unfavorable discharge disposition and higher mortality risk compared to falling levels. Falling troponin levels were more likely to be isolated in the absence of ECG or echocardiographic changes compared to rising levels (53% vs. 22%).^[68]

3.1.2. Left ventricular dysfunction

Acute ischemic stroke has also been associated with left ventricular dysfunction. In a study of 644 patients with ischemic stroke, 9.6% of these patients had systolic dysfunction, 23.3% had diastolic dysfunction without systolic dysfunction, and 5.4% had clinically overt heart failure.^[69] An even higher prevalence was reported with the use of cardiac magnetic resonance imaging (CMR), where systolic dysfunction was observed in 25% and diastolic dysfunction in 59% of patients with ischemic stroke without known AF.^[70] Left ventricular dysfunction itself is a risk factor for ischemic stroke,^[71] and nearly 5% of patients with acute ischemic stroke who are in sinus rhythm have a depressed left ventricular ejection fraction (LVEF).^[72] The underlying mechanism is thought to be due to increased blood stasis within the left ventricular and left atrium, leading to intracardiac thrombus formation and cardioembolic stroke.^[71] It is therefore difficult to determine if the association between stroke and left ventricular dysfunction is due to neurogenic effects on the heart in the stroke-heart syndrome or underlying undiagnosed left ventricular dysfunction, predisposing the individual to stroke in the first place.

Experimental animal models have shown that induced cerebral artery occlusion produces left ventricular dysfunction accompanied by elevated cardiac enzymes.^[44,45] In mice, greater severity of right hemispheric atrophy following induced stroke was found to correlate with lower LVEF, and beta-blockade appeared to decelerate extracellular cardiac remodeling.^[49] Clinical data in humans also appear to support a causal relationship between stroke and myocardial dysfunction. A study of 213 patients with first-ever acute ischemic stroke without pre-existing heart disease showed that patients with impaired LVEF had suffered from strokes with higher infarction volumes than those with preserved LVEF. In addition, multivariate voxelwise lesion analysis found associations between decreased LVEF and damaged voxels in specific areas of the brain, including the insula and amygdala of the right hemisphere.^[73] Another study showed that acute ischemic stroke affecting the right insula and peri-insular regions, as well as the left parietal cortex, were associated with impaired left ventricular global longitudinal strain.^[74] These findings are consistent with the prominent role of certain brain areas, such as the right hemisphere in autonomic regulation within the heart-brain axis, and further support the hypothesis of the stroke-heart syndrome. Echocardiography or CMR may help distinguish certain cases of post-stroke myocardial dysfunction from pre-existing myocardial dysfunction with resultant cardioembolic stroke, such as the identification of an intracardiac thrombus.^[75] CMR performed in patients with acute ischemic stroke showed focal myocardial fibrosis in 34% of patients, of which 74% showed an ischemic pattern. Similar proportions of focal fibrosis were seen even in patients without known prior myocardial infarction.^[76] Interval imaging showing resolution of myocardial dysfunction supports a neurogenic etiology, which is more likely to be transient.

3.1.3. AF and other arrhythmias

Cardiac arrhythmias are associated with acute ischemic stroke and occur more frequently in patients with cerebral hemisphere infarctions compared to brain stem infarctions.^[77] A prospective study of 501 patients, which assessed telemetry data during the first 72 h of acute stroke, showed that significant cardiac arrhythmias occurred in 25% of these patients.^[40] Tachyarrhythmias (AF with rapid ventricular response, focal atrial tachycardia, non-sustained ventricular tachycardia, etc) were more common than bradyarrhythmias (AF with slow ventricular response, atrioventricular blocks, sinus node dysfunction, and asystole). The development of arrhythmias was independently associated with neurological deficits of greater severity, as measured by the National Institutes of Health Stroke Scale on admission, and arrhythmia risk was the highest in the first 24 h and declined over 3 d.^[40] The relationship between arrhythmias and the acuity and severity of neurological insults provides supportive evidence of a neurogenic etiology. Abnormalities on the 12-lead ECG, such as ST-T changes and prolonged QT interval, were also frequently observed in patients with acute ischemic stroke.^[78]

AF is a well-known risk factor for cardioembolic stroke (see section 4.1.1),^[79,80] and it is therefore unsurprising to identify a high incidence of AF in patients with acute ischemic stroke or transient ischemic attack.^[81] However, the autonomic imbalances and catecholamine surge in acute stroke itself may be involved in triggering new-onset AF. In 1993, a study of patients with first-ever stroke and new AF showed that AF resolved spontaneously after a few days in more than half of patients, leading the authors to conclude that such instances of AF may have occurred as a consequence of stroke.^[82] AF detected after stroke (AFDAS) is defined as newly-detected AF after ischemic stroke or transient ischemic attack in patients without known AF.^[83] In contrast to patients with known AF, those with AFDAS have been shown to have a lower prevalence of coronary artery disease, congestive heart failure, prior myocardial infarction, and a history of cerebrovascular events.^[84]

In a study of 275 patients with acute ischemic stroke, those with AFDAS also had a lower prevalence of underlying structural abnormalities such as left atrium enlargement, and these patients had a higher proportion of insular involvement than patients with known AF. Larger infarctions of 15 mm or greater were more common in patients with AFDAS compared to those in sinus rhythm, suggesting a relationship between AF development and stroke severity and location.^[9] Another study showed that AF was more frequently detected in strokes with insular cortex involvement and higher levels of cardiac enzymes.^[85] In the Rate of Atrial Fibrillation Through 12 Months in Patients With Recent Ischemic Stroke of Presumed Known Origin (STROKE-AF) study, AFDAS detected via implantable loop recorders (ILRs) was found in 12.1% of patients with strokes caused by large- or small-vessel disease at 12 months.^[86] This further increased to 21.7% of patients after continued monitoring for 3 years.^[87] This suggests that there is a high detection rate of AFDAS even in cases where AF causing cardioembolic stroke is unlikely. Analogous to clinical studies, experimental animal models support neurogenic mechanisms in the pathogenesis of AFDAS. Induced left and right insular ischemic strokes in rats were found to produce left atrium endothelial dysfunction, myocardial inflammatory infiltration, and fibrosis. The extent of tissue scarring also correlated with the severity of stroke-related microglia activation.^[88] These observations suggest acute stroke may trigger AF by increasing the left atrium substrate for AF development. There exist prognostic differences between AFDAS and known AF, with a meta-analysis reporting a 26% lower risk of recurrent stroke in AFDAS compared to known AF.^[84] Management considerations for AF in patients with acute stroke are discussed in greater detail in section 4.1.1.

3.1.4. Post-stroke neurogenic sudden cardiac death

The exact mechanism of sudden cardiac death after stroke is not well understood. It may be related to autonomic imbalances and interactions between the initial neurological insult and cardiovascular complications such as myocardial injury and infarction, myocardial dysfunction, and arrhythmias.^[89] Patients with early severe cardiac complications after stroke have a 2- to 3-fold increased risk of short-term mortality.^[60,90] Neurogenic sudden cardiac death is discussed in greater detail in section 3.4.

3.2. Neurogenic pulmonary edema and cardiomyopathy

Neurogenic pulmonary edema presents as acute respiratory distress arising in the context of severe neurological injury, such as ischemic and haemorrhagic stroke, traumatic brain injury, and seizures.^[91,92] Its pathophysiology is thought to involve sympathetic hyperstimulation and excess catecholamine release in response to abrupt increases in intracranial pressure, resulting in both pulmonary and systemic vasoconstriction. Elevated pulmonary vascular pressure alters Starling forces and leads to extravasation of fluid into the lung interstitium. Both raised pulmonary and systemic vascular resistance lead to increased myocardial demands and may cause reversible myocardial stunning or injury, such as in Takotsubo syndrome. A case series of 5 women with subarachnoid hemorrhage (SAH) and neurogenic pulmonary edema showed echocardiographic evidence of reduced global and segmental left ventricular systolic function. Interval echocardiography 2 to 6 weeks after the acute intracranial hemorrhage (ICH) later showed normal left ventricular function in all 3 of the patients who survived.^[93] Severity of cardiac dysfunction in association with ICH is directly related to the degree or acuity of ICH,^[94] and has been shown to improve over a week in parallel with the normalization of catecholamine concentrations.^[95]

3.2.1. Takotsubo syndrome

Takotsubo cardiomyopathy is a type of neurogenic cardiomyopathy and was first described by Kurisu et al^[96] as a transient “tako-tsubo (octopus trap)-like left ventricular dysfunction” mimicking acute myocardial infarction, and which subsequently resolved on left ventricular ventriculography about a week later. Patients typically present with acute onset chest pain or dyspnoea and ECG changes are easily mistaken for a myocardial infarction, including ST segment elevations. In severe cases, it may be complicated by severe heart failure, cardiogenic shock, or malignant arrhythmias.^[97] There are several subtypes of Takotsubo cardiomyopathy: (1) apical and midventricular dyskinesis, akinesis, or hypokinesis with basal sparing, giving rise to the classical apical ballooning pattern (most common subtype); (2) circumferential midventricular wall motion abnormality with basal and apical hyperkinesis; (3) reverse Takotsubo syndrome with basal involvement and mid segment and apical sparing; (4) focal left ventricular involvement; and (5) isolated right ventricular involvement.^[98,99] Takotsubo syndrome was classically described in the context of emotional stress, such as an unexpected bereavement, conflict, or major life event, and was hence termed “broken heart syndrome”. It is now known that a clear preceding stressor is not always identified, and Takotsubo syndrome can also occur during physical stress, including neurological conditions such as acute ischemic stroke,^[10,100,101] ICH,^[101–103] and epilepsy.^[101] Takotsubo cardiomyopathy in patients with ischemic stroke is often observed in relation to infarctions that include the insular cortex or peri-insular areas.^[10,100] A previous study identified 23 patients with specific echocardiographic findings consistent with Takotsubo cardiomyopathy from a multi-center stroke registry database. These patients were predominantly older women, and 91.3% had a typical apical ballooning pattern of myocardial dysfunction. Takotsubo syndrome in these patients was associated with neurological deterioration, poor functional outcomes, and high mortality.^[100]

Exaggerated sympathetic stimulation from acute stress is believed to be central to the pathophysiology of stress-induced cardiomyopathy. In an in vivo rat model, administration of an intravenous epinephrine bolus produced the characteristic reversible apical depression of myocardial contraction together with basal hypercontractility.^[104] The study also showed a greater number of beta-2 adrenergic receptors and increased functional response in apical compared to basal cardiomyocytes, supporting the observation of higher apical sensitivity to circulating epinephrine.^[104] Another study reported significantly higher concentrations of norepinephrine, epinephrine, and dopamine in the atria compared to the ventricles of the canine heart. In particular, concentrations were lower in the left ventricular apex compared to the base.^[105] These findings may explain why the majority of Takotsubo cardiomyopathy cases involve apical dysfunction with sparing of the base, and they also lend support to the catecholaminergic pathogenesis of the disease.

Similar observations have been made in clinical studies. Patients with left ventricular dysfunction after sudden emotional stress had significantly elevated plasma catecholamine levels, which were markedly higher than among those with Killip class Ⅲ myocardial infarction.^[54] A case series reported left ventricular ballooning and a median ejection fraction of 35% following intravenous administration of epinephrine or dobutamine, with recovery of left ventricular systolic function after a median of 7 d.^[106] Furthermore, cocaine use, which is known to inhibit the re-uptake of catecholamines in presynaptic neurons, has been reported in association with apical ballooning of the left ventricular seen on left ventricular ventriculography.^[107] These studies provide further evidence of the link between catecholamine excess and Takotsubo syndrome. Pharmacological therapy in Takotsubo syndrome is heterogenous and often extrapolated from treatment strategies established in the context of myocardial infarction. A recent case-control study of 620 patients showed that conventional cardiovascular medications were associated with better survival in myocardial infarction patients but not in patients with Takotsubo syndrome.^[108] Management of cardiogenic shock in Takotsubo syndrome is also challenging, as most inotropic agents potentially increase the catecholamine surge and, thus, worsen left ventricular systolic dysfunction. Expert consensus recommends mechanical haemodynamic support, such as intra-aortic balloon counterpulsation, temporary left ventricular assist devices, and extracorporeal membrane oxygenation, in the management of these patients.^[109]

3.3. Electrocardiographic changes and arrhythmias in intracranial haemorrhage

There is a high incidence of ECG changes in patients with ICH.^[110,111] These ECG abnormalities include QT interval prolongation, ST-T morphologic changes, sinus bradycardia, and inverted T waves.^[111–114] QT prolongation in particular is often associated with ICH involving the insular cortex, intraventricular hemorrhage, and hydrocephalus.^[111,112] Less commonly, ST-segment elevation has also been described in patients with SAH in the absence of coronary artery stenosis on invasive coronary angiography.^[115] Both tachy- and brady-arrhythmias are common in patients with SAH, including ventricular premature complexes, non-sustained ventricular tachycardia, supraventricular premature complexes, paroxysmal supraventricular tachycardia or AF, sinoatrial block or sinus arrest, atrioventricular blocks, and idioventricular rhythms.^[116] Frequency and severity of arrhythmias appear to be highest within the first 48 h of acute SAH. Life-threatening ventricular arrhythmias such as Torsades de Pointes degenerating into ventricular fibrillation in patients with SAH were associated with QT interval prolongation.^[116]

Although the ECG patterns observed in acute ICH are well described, their clinical significance is less known. Several studies have reported echocardiographic abnormalities in association with acute ECG changes in patients with ICH.^[115,117] ST-segment elevation in SAH without organic stenosis or vasospasm on angiography was associated with transient corresponding regional wall motion abnormalities on echocardiography.^[115] Another study showed that symmetrical T wave inversions and severe QT interval prolongation of 500 milliseconds or longer occurred more frequently in patients with SAH with echocardiographically-detected wall motion abnormalities.^[117] The underlying pathophysiology of the ECG changes seen in patients with acute cerebrovascular disorders may be autonomic neural stimulation from the hypothalamus or elevated circulating catecholamines, resulting in neurogenic stunned myocardium.

A small case series, however, reported that out of 45 patients with SAH or intracranial aneurysms, 4 patients had wall motion abnormalities but only minor ECG changes, whereas other patients with classical ECG patterns such as deep inverted T waves had normal echocardiograms. The patients who had wall motion abnormalities had severe neurological dysfunction, and the authors concluded that myocardial dysfunction might be more closely related to severity of the underlying neurological condition rather than ECG features.^[118] In terms of prognosis, certain ECG changes in ICH, such as QT prolongation and non-specific ST-T changes, confer a high mortality rate.^{[114,119,120]} Patients with evolving ECG changes were found to have worse outcomes compared to those whose ECGs were either consistently normal or consistently abnormal.^[110] Serial ECG recording in 61 patients with aneurysmal SAH showed that ECG changes correlated with poorer prognosis overall but did not correlate with specific outcome events including cardiac disease. Instead, ECG changes were associated with the initial level of consciousness.^[119] Another study showed that among 58 patients with SAH and ECG changes, no deaths from cardiac causes occurred, and ECG changes were associated with more severe neurological injury.^[121] These findings suggest that the ECG abnormalities in ICH may not signify impending cardiac complications but may instead reflect adverse intracranial status. Therefore, treatment of the underlying brain injury should be the focus of management in these patients.

3.4. Sudden cardiac death

Cardiac arrest or sudden cardiac death from neurological disease is uncommon and can occur in ICH, epileptic seizures, and ischemic stroke.^[122–126] This may be a result of: (1) a dramatic catecholamine surge leading to myocardial stunning, and (2) abrupt raised intracranial pressure with brain stem dysfunction and respiratory arrest causing hypoxia, which then activates adenosine release, leading to decreased cardiac contractility, atrioventricular nodal conduction, and automaticity.^[124] These patients have a very poor prognosis.^[126]

3.4.1. Sudden unexpected death in epilepsy (SUDEP)

SUDEP is defined as a sudden, unexpected, witnessed or unwitnessed, non-traumatic and non-drowning death in patients with epilepsy with or without evidence of a seizure, in which post-mortem examination does not reveal a structural or toxicologic cause for death.^[127] Cardiac dysfunction is thought to be integral to the pathophysiology of SUDEP. Epilepsy may induce rhythm abnormalities such as ictal asystole, post-ictal asystole, ictal bradycardia, ictal atrioventricular block, post-ictal atrioventricular block, post-ictal AF or atrial flutter, and post-ictal ventricular fibrillation.^[128–130] The left hemisphere appears to play a preferential role in the genesis of ictal bradycardia,^[129] further supporting the role of the left hemisphere in parasympathetic control of the heart.^[17] These post-ictal arrhythmias were often associated with SUDEP.^[130]De novo mutations, previously reported pathogenic mutations, or candidate pathogenic variants in cardiac arrhythmia and epilepsy genes have been described in SUDEP cases, suggesting that there may be a common genetic predisposition.^[131] Structural pathology has also been found in association with SUDEP, and post-mortem studies revealed a large proportion of cardiac and pulmonary abnormalities such as myocyte hypertrophy and myocardial fibrosis of various degrees, as well as pulmonary congestion in SUDEP cases.^[132] Takotsubo cardiomyopathy and neurogenic stunned myocardium have also been described.^[132]

3.4.2. Catecholaminergic polymorphic ventricular tachycardia (CPVT)

CPVT is a rare arrhythmogenic disorder in children and adolescents, with a high mortality rate.^[133] As its name suggests, CPVT is characterized by adrenergic-induced polymorphic ventricular tachycardia, classically bidirectional.^[134] Specific genetic mutations have been identified in different CPVT subtypes.^[135–138] The ECG in patients with CPVT is typically normal at rest, with progressive ventricular ectopy seen on exercise testing. Ventricular premature complexes usually have a right bundle branch block morphology with alternating right and left axis deviation. With continued exercise, these progress to non-sustained and then sustained polymorphic bidirectional ventricular tachycardia, which can lead to cardiogenic syncope. In some cases, CPVT degenerates into ventricular fibrillation and, consequently, sudden cardiac death.^[139] CPVT is triggered by a catecholaminergic surge, such as in exercise and stress testing, isoproterenol infusion, and extreme emotion. Accordingly, beta-blockade is the first-line treatment and non-selective beta-blockers such as nadolol or propranolol are preferred.^[140]

4. Neurological effects of cardiac disease

4.1. Cardioembolic strokes

Cardioembolic strokes make up an estimated 25% of all ischemic strokes, of which the most common is AF-associated cardioembolic stroke. Other sources of cardioembolism include left ventricular thrombus in patients who have suffered a myocardial infarction or with cardiomyopathy or valvular heart disease.^[79] Congenital cardiac abnormalities such as patent foramen ovale are not strictly a cause of cardioembolic stroke per se, but they may provide a conduit for paradoxical embolization.

4.1.1. AF-related stroke

AF confers a 4-fold to 5-fold increase in stroke risk,^[80] and AF-related ischemic stroke has been associated with increased stroke severity and a mortality rate nearly twice that of non-AF stroke.^[141] These patients have also been shown to be at higher risk for recurrent strokes.^[141] Among patients with ischemic stroke, extended monitoring using ILRs was associated with increased AF detection rates in the Post-Embolic Rhythm Detection With Implantable Versus External Monitoring (PERDIEM) trial.^[142] The STROKE-AF trial also showed that the use of implantable cardiac monitors (ICMs) in patients with ischemic stroke detected significantly more AF compared to usual care.^[86] The rate of AF detection in STROKE-AF remained higher in patients with ICM compared to controls after 3 years.^[87] However, the clinical significance of such detected AF episodes is uncertain. The Impact of Standardized MONitoring for Detection of Atrial Fibrillation in Ischemic Stroke (MonDAFIS) trial showed that systematic Holter-ECG monitoring up to 7 d in the inpatient setting in patients with acute ischemic stroke increased AF detection but had no effect on rate of oral anticoagulation (OAC) use or cardiovascular outcomes and death when compared to standard diagnostic procedures.^[143] The ongoing Intensive Rhythm Monitoring to Decrease Ischemic Stroke and Systemic Embolism Study (Find-AF 2) randomized, controlled, open-label parallel group trial aims to study patients with symptomatic ischemic stroke without known AF and without detected AF on an additional 24-hour Holter-ECG. These patients are randomized to either prolonged, intensified ECG monitoring or standard care. In the intervention arm, patients with a high risk of underlying AF receive an ICM whereas those not at high risk receive repeated 7-day Holter-ECGs. The primary endpoint of the study is the time to recurrent ischemic stroke or systemic embolism.^[144] This trial may help clarify the clinical significance of AF detected on intensive post-ischemic stroke AF screening. Other trials have evaluated methods of AF screening in patients with stroke risk factors. The Atrial Fibrillation Detected by Continuous ECG Monitoring (LOOP) study showed that in patients without AF but with at least 1 additional stroke risk factor, ILR use led to increased AF detection and OAC initiation but did not affect risk of stroke or systemic embolism.^[145] A dedicated AF screening program may be beneficial in select patients, however, as shown in the Systematic ECG Screening for Atrial Fibrillation Among 75 Year Old Subjects in the Region of Stockholm and Halland, Sweden (STROKESTOP) trial, which showed a small decrease (hazard ratio: 0.96, confidence interval: 0.92–1.00) in the primary endpoint of stroke, systemic embolism, severe bleeding, and all-cause death in patients who were offered AF screening.^[146]

Subclinical AF is characterized by brief asymptomatic episodes of AF lasting between 6 min and 24 h, detected on continuous rhythm monitoring in patients with a pacemaker, defibrillator, or ICM. Some have proposed that subclinical AF should not include individuals with AF detected after ischemic stroke, as, despite the absence of AF-specific symptoms, these patients have already suffered from embolic complications of AF and should not be regarded as “asymptomatic”.^[83] Currently, quantification of stroke risk in subclinical AF remains a clinical challenge. Patients with subclinical AF were reported to have a 2.5-fold risk of stroke or systemic embolism,^[147] although this risk may be lower in those with a shorter duration of AF.^[148,149] While anticoagulation is the treatment of choice for primary and secondary stroke prevention in clinical AF in the presence of stroke risk factors,^[150] the role of OACs in subclinical AF and other atrial high-rate episodes is less established. The Apixaban for the Reduction of Thrombo-Embolism in Patients With Device-Detected Sub-Clinical Atrial Fibrillation (ARTESiA) trial randomized 4,012 patients with subclinical AF detected on ICMs to apixaban vs. low dose aspirin. The study showed that apixaban use resulted in a lower risk of stroke or systemic embolism but an increased risk of major bleeding as compared to low dose aspirin.^[151] However, the absolute risk reduction in stroke with apixaban use was low, at 1.6% over 3.5 years, accompanied by a similar increased rate of major bleeding. The Non-vitamin K Antagonist Oral Anticoagulants in Patients With Atrial High Rate Episodes (NOAH-AFNET 6) trial, which randomized 2,600 older patients with atrial high-rate episodes and at least 1 other risk factor for stroke to edoxaban vs. aspirin or placebo, depending on clinical indication, found no difference in outcomes of cardiovascular-related death, stroke, or systemic embolism between the 2 groups. The edoxaban group, however, had increased bleeding complications.^[152] A subsequent study-level meta-analysis of ARTESiA and NOAH-AFNET 6 concluded that, overall, OAC reduced stroke risk in patients with subclinical AF at the expense of increased risk of major bleeding.^[153] Ultimately, the decision whether or not to anticoagulate these patients will depend on the individual’s preference and risk profile.

Stasis of blood flow and thrombus formation within the left atrium appendage is thought to be the main source of cardioembolism in AF, and there is growing interest in left atrial appendage occlusion as a method of stroke risk reduction in patients with AF.^[154] Overall, data from key trials like WATCHMAN Left Atrial Appendage System for Embolic PROTECTion in Patients With Atrial Fibrillation (PROTECT AF),^[155] Evaluation of the WATCHMAN Left Atrial Appendage Closure Device in Patients With Atrial Fibrillation Versus Long Term Warfarin Therapy (PREVAIL),^[156] and Left Atrial Appendage Closure vs. Novel Anticoagulation Agents in Atrial Fibrillation (PRAGUE-17) suggest that left atrial appendage occlusion is non-inferior to anticoagulation in carefully selected patients.^[157] However, non-inferiority in stroke prevention appears to be driven by lower rates of hemorrhagic stroke compared to anticoagulation, and, indeed, the rate of ischemic stroke or systemic embolism was higher in the device arm in PROTECT AF and PREVAIL.^[158] Surgical left atrium appendage closure during planned cardiac surgery has also been explored with promising results.^[159,160]

There is no universally accepted definition for cryptogenic stroke; however, most are in agreement that it is a diagnosis of exclusion where there is no identifiable source of cardioembolism, large- or small-vessel disease after extensive evaluation.^[161–163] However, there is evidence that the majority of cryptogenic stroke has an embolic pathogenesis. Cortical infarctions, which are suggestive of embolism, are common in cryptogenic stroke,^[164] and a previous study reported that nearly two-thirds of patients initially diagnosed with cryptogenic stroke were later found to have potential sources of cardioembolism.^[165] There is currently no established management strategy for patients with cryptogenic stroke. The recent AtRial Cardiopathy and Antithrombotic Drugs In Prevention After Cryptogenic Stroke (ARCADIA) trial, which randomized 1,015 patients with cryptogenic stroke and evidence of atrial cardiopathy without AF to apixaban or aspirin, showed that apixaban did not significantly reduce recurrent stroke risk compared to aspirin.^[166]

4.1.2. Left ventricular thrombus

Left ventricular thrombus occurs in acute myocardial infarction as well as in ischemic and non-ischemic cardiomyopathies. Stasis of blood within the left ventricular cavity due to regional wall akinesia and dyskinesia, together with subendocardial injury and inflammation—especially in the post-myocardial infarction state—and hypercoagulability contribute to left ventricular thrombus formation.^[167,168] Systemic embolic events are known complications of left ventricular thrombus, of which the most serious is acute ischemic stroke.^[169–171] Ischemic stroke has been reported in up to 15% of patients with left ventricular thrombus, although the true incidence of stroke including silent strokes is unknown as neuroimaging is not routinely performed in these patients.^{[169,171,172]} Anticoagulation is the treatment of choice for stroke prevention in left ventricular thrombus, and guidelines recommend warfarin or a direct oral anticoagulant for 3 to 6 months, with anticoagulation duration guided by interval echocardiography or CMR.^[173] The optimal duration of anticoagulation for left ventricular thrombus is unknown, but prolonged anticoagulation may be considered in the presence of high-risk thrombus characteristics, persistence of depressed LVEF and myocardial akinesia or dyskinesia, or a persistent prothrombotic state.^[174]

4.2. Cognitive impairment in cardiac disease

Cognitive impairment is common in patients with cardiovascular disease and may be due to a combination of shared risk factors and direct consequences of cardiac dysfunction on the brain [Figure 4].^[175] Vascular risk factors such as high body mass index, smoking, hypertension, diabetes, and hyperlipidemia have been associated with elevated amyloid uptake on positron emission tomography imaging, consistent with the role of vascular disorders in the development of Alzheimer disease.^[176,177] Hypertension has also been linked to an increased risk of Alzheimer disease, and this risk increase was more pronounced in stage 2 compared to stage 1 systolic hypertension.^[178] Patients with familial hypercholesterolemia were found to have a higher incidence of cognitive impairment compared to those without the disease.^[179] Beta-amyloid proteins may also be implicated in both Alzheimer disease and vascular inflammation. Beta-amyloid was associated with progression of arterial stiffness, incident subclinical atherosclerosis, and coronary artery disease.^[180] Cardiac disease itself may also play a role in the development of cognitive impairment. N-terminal pro B-type natriuretic peptide and high sensitivity cardiac troponin levels have been associated with beta-amyloid proteins as well as increased dementia risk,^[181–183] independent of traditional cardiovascular risk factors.

Figure 4.

Potential pathological mechanisms linking cardiac disease and cognitive impairment.

4.2.1. Cognitive impairment in heart failure

Cognitive impairment and dementia are common in patients with heart failure and may be seen in both older and younger patients.^[184,185] Development of cognitive deficits in these patients is associated with functional decline,^[186] increased re-hospitalizations,^[187–189] and increased mortality.^[188–190] Although heart failure and cognitive impairment have common risk factors, studies that included a control group with cardiovascular disease without heart failure have shown that cognitive impairment remains higher in the heart failure group.^[191,192] The prevalence of cognitive impairment among patients with incident heart failure was found to be similar to controls without heart failure, and it was lower than reported in the general heart failure population. This suggests that cognitive impairment may develop after the onset of heart failure and may be a consequence of the disease.^[193] Modified Mini-Mental State Examination scores also declined over 5 years in patients with incident heart failure compared to controls.^[194] Furthermore, studies have shown a dose-response relationship between heart failure severity and cognitive impairment. Patients with more advanced New York Heart Association class disease had a greater incidence of cognitive impairment,^[195,196] and they had poorer memory, visuospatial ability, psychomotor speed, and executive function.^[197,198] Cognitive impairment in patients with acute decompensated heart failure improved after compensation.^[199] Potential pathophysiological pathways in heart failure include cerebral hypoperfusion and systemic inflammation [Figure 4]. Low ejection fraction has been associated with lower cognitive scores,^[200,201] and this is believed to be due to cerebral hypoperfusion. Patients with heart failure were found to have lower cerebral blood flow velocity,^[202–204] which was associated with impaired cognitive function.^[205,206] Cerebral hypoperfusion may lead to cognitive impairment via occult infarction and development of white matter lesions as well as reduced cerebral metabolism with resultant gray matter atrophy.^{[185,202,207,208]} Heart failure also induces a systemic inflammatory state with elevated levels of circulating cytokines. It is postulated that these cytokines are involved in neuronal degeneration in patients with heart failure.^[185,209] Screening for cognitive impairment in patients with heart failure is important. The Mini-Mental State Examination or the Montreal Cognitive Assessment are quick screening tools that have been found to be sensitive and specific when compared against a gold standard neuropsychological test battery, and they can be easily performed in the outpatient setting.^[210] Some studies have shown improved cognition in patients after the initiation of pharmacological or device therapy for heart failure.^[211,212]

4.2.2. Aortic stenosis and cognitive impairment

Severe aortic stenosis has been associated with cognitive impairment.^[213–215] Apart from shared risk factors and the burden of aortic stenosis in older patients, severe aortic stenosis may contribute to cognitive decline via reduced cardiac output and cerebral blood flow. There are studies which report cognitive improvement following transcatheter aortic valve intervention or surgical aortic valve replacement,^[214–216] possibly due to post-intervention increased cerebral blood flow.^[217,218] However, transcatheter aortic valve intervention and surgical aortic valve replacement may also pose a risk of acute silent ischemic cerebral lesions via microembolism,^[219,220] and a systematic review of 6 studies showed that, overall, there was no difference in cognitive function before and after valve intervention.^[221]

5. Future directions

Despite increased recognition of the heart-brain axis, there remain uncertainties in the pathophysiology, diagnosis, management, and prognosis of neurocardiogenic conditions. In the stroke-heart syndrome, an elevated cardiac troponin level is common, but its diagnostic and therapeutic implications are unclear. The ongoing PRediction of Acute Coronary Syndrome in Acute Ischemic StrokE (PRAISE) trial, where patients with stroke and elevated troponin levels undergo invasive coronary angiography, aims to study whether dynamic troponin levels (ie, rise or fall pattern) indicate the presence of acute coronary syndromes as compared to chronic troponin elevation.^[222] This may provide clarity on the pathophysiology and significance of raised troponins in stroke and thus guide clinical decision making. CMR may offer better accuracy in detecting subtle differences associated with chronic cardiac or neurogenic ischemia,^[223] and the ongoing prospective observational Cardiomyocyte Injury Following Acute Ischemic Stroke (CORONA-IS) study will investigate the role of CMR in differentiating mechanisms of post-stroke myocardial injury.^[224] In patients with Takotsubo syndrome, the Optimized Pharmacological Treatment for Broken Heart (Takotsubo) Syndrome (BROKEN-SWEDEHEART) study aims to investigate the roles of adenosine and dipyridamole in accelerating cardiac recovery, as well as that of apixaban in preventing thromboembolic complications.^[225] This will provide new data on therapeutic options in a condition where most of the management is currently supportive. Anticoagulation for AF in patients with a high bleeding risk is challenging, and this includes patients with newly detected AF following ICH. The Anticoagulation in ICH Survivors for Stroke Prevention and Recovery (ASPIRE),^[226] EdoxabaN for IntraCranial Hemorrhage Survivors with Atrial Fibrillation (ENRICH-AF),^[227] and PREvention of STroke in Intracerebral haemorrhaGE Survivors With Atrial Fibrillation (PRESTIGE-AF)^[228] trials aim to study the efficacy and safety of anticoagulation for cardioembolic stroke prevention with direct oral anticoagulants vs. antiplatelets vs. no anti-thrombotics for stroke risk reduction in patients at high risk with previous ICH. For heart failure patients with cognitive impairment, there is limited randomized data on whether guideline-directed medical therapy can improve cognitive function, and this can be explored in future studies.

6. Conclusions

Our understanding of the interactions between the heart and the brain has progressed tremendously since Cannon’s preliminary observations in 1942.^[1] Recognition and characterization of the stroke-heart syndrome allow clinicians to identify patients at risk of developing serious post-stroke cardiac complications (eg, myocardial dysfunction and arrhythmias), as well as develop appropriate diagnostic approaches to conditions such as neurogenic AF. Patients with heart failure may benefit from screening for cognitive impairment, and further investigations still are needed to clarify anticoagulation strategies in those with left ventricular thrombus or subclinical AF for stroke prevention. Interdisciplinary collaborations among cardiologists, neurologists, cardiac surgeons, neurosurgeons, and acute care physicians are needed to provide comprehensive care for these patients and to advance the field of neurocardiology research.

Funding

Ching-Hui Sia was supported by the Singapore Ministry of Health’s National Medical Research Council under its Clinical Scientist Individual Research Grant New Investigator Grant (NMRC/MOH-001080) and the Transition Award (NMRC/MOH001368-00).

Author contributions

Fang Qin Goh participated in writing the manuscript. Benjamin Y.Q. Tan, Leonard L.L. Yeo, and Ching-Hui Sia contributed to reviewing and improving the manuscript. Ching-Hui Sia supervised the writing of the manuscript. All authors read and approved the final manuscript.

Conflicts of interest

None.