Heart and brain interactions

Several studies, including large population-based analyses, report associations between MHD and incident CVD [11]. Bi-directional interrelations between major depression and CVD suggest that both diseases may share common pathophysiological pathways [12]. These include lifestyle factors (e.g. physical activity, smoking behaviour), dysfunction of endocrine systems (e.g. hypothalamus–pituitary–adrenal axis), and an imbalance of pro- and anti-inflammatory factors. Both MHD and CVD share genes involved in energy metabolism, stress system, circadian rhythm, inflammation and neurotransmission. For example, covariation of depressive symptoms and coronary disease may in part be attributable to common genetic vulnerability potentially related to inflammation, serotonin pathways or mitochondrial energy metabolism [13, 14]. Associations between a functional sequence variant of the neuropeptide S receptor‑1 gene, which regulates anxiety, and clinical outcomes and healthcare utilization in HF patients suggests the possibility that psychogenetic determinants may modulate also such endpoints [15]. Figure 1 shows a simplified schematic of key heart–brain interactions, which may induce systemic organ dysfunction and promote development and progression of different clinical phenotypes of both CVD and MHD. The pathogenetic importance of each factor varies in individual patients, and likely depends on the overall risk profile and personal resilience (i.e. the capacity to adapt swiftly and successfully in the face of physical and/or emotional challenges), plus the presence and severity of the somatic illness [16].

Similar to external stress, negative emotions might impact adversely on neurohormonal regulatory circuits. Autonomic nervous dysfunction may trigger multiple biological changes, including increased sympathetic tone, innate and adaptive immune system activation, higher circulating levels of stress hormones (e.g. cortisol and pro-inflammatory cytokines) and metabolic dysregulation [17]. Systemic inflammation may induce a procoagulatory state and endothelial dysfunction/injury. All these contribute to the development and progression of atherosclerosis, and increase the risk for arrhythmias such as AF, cardiac and cerebral clinical events and systemic disease in general [12, 18]. For example, recent evidence suggests that epicardial adipose tissue is a direct source of pro-inflammatory cytokines and could mediate various deleterious effects of systemic inflammation on cardiac structure and function, thus contributing to the pathogenesis of CVD, including calcific aortic stenosis [19]. Interestingly, transcatheter aortic valve replacement was shown to persistently improve depression and anxiety in such patients [20]. Other studies have demonstrated increased intra-abdominal and epicardial adipose tissue in individuals who are physically healthy but depressed, and in cardiac patients (e.g. congenital heart disease) with major depressive disorder [14, 21]. Lastly, MHD may influence behavioural factors such as lifestyle, self-care and adherence to evidence-based therapies [22], which further contributes to a vicious circle of declining health, functional capacity and personal socioeconomic status.

Mental health disorders can be easily overlooked or misinterpreted by physicians and patients [6] because signs and symptoms such as physical disability/fatigue, loss of interest, anhedonia and sad mood, or eating and sleep disorders may be associated with both MHD and CVD, particularly at more advanced disease stages. Systematic screening for MHD using simple validated self-report tools offers the potential for early identification and targeted management, and is widely recommended [5]. Screening for MHD in cardiologist practices and hospitals would require minimal resources and efforts, but identification of MHD might necessitate increased downstream support from mental health providers, including immediate evaluation of suspected suicidality, and thus impact significantly on routine CVD management algorithms [5]. This could be one reason why MHD screening has not been widely implemented in CVD care settings, and why MHD remain underdiagnosed and undertreated. Additional research is needed to better determine the benefit of screening CVD patients for MHD [23]. Guideline-supported screening algorithms for MHD, especially depression, and additional diagnostic steps needed after a positive screening result are detailed in the work of Jha et al. [5].

Given the close correlation between CVD severity and the prevalence and severity of MHD, especially depression [12], treatments that improve cardiovascular function should also benefit coexisting mood disorders. The recent MEMS-HF study evaluating haemodynamic-guided HF management based on remote pulmonary artery pressure (PAP) monitoring demonstrated that decreases in PAP were associated with significant, persistent remission of depressive symptoms, and that larger PAP decreases resulted in greater depression remission and quality of life improvement [24]. This is the first time that an MHD (depression) has been linked with a treatable biological variable (haemodynamic congestion). Accordingly, standard care approaches should include disease-modifying treatments for CVD and associated comorbidities as well as MHD, and integrate needs-adjusted management of stressful signs and symptoms (e.g. sleep optimization, pain management) while improving patients’ psychosocial functioning, healthcare competence and self-monitoring/empowerment [12].

Specific antidepressant and anxiolytic treatments, including pharmacotherapy, psychological or behavioural interventions and/or physical exercise, have shown disparate effects on symptoms and clinical outcomes in patients with CVD, including post-stroke depression [25, 26]. Due to their overall safety profile and effectiveness, selective serotonin re-uptake inhibitors (SSRI) have emerged as first-line therapy for depression and anxiety in patients with CVD [25]. Evidence suggests that SSRI are safe in patients with stable CVD and after myocardial infarction, and may improve depressive symptoms and possibly prognosis (overview in [5]). However, SSRI had no effect on either depression or outcomes compared with clinical management alone in patients with symptomatic HF in two large randomized trials [27, 28], and dose-dependent QTc prolongation indicates increased risk for torsade-de-pointes arrhythmias [29]. Meta-analysis data suggest that SSRI may increase mortality risk in HF patients [30]. Few data are available on the safety and efficacy of newer antidepressants such as melatonergic agonists (agomelatine) and norepinephrine and dopamine reuptake inhibitors (bupropion) in patients with CVD. Adverse clinical effects such as changes in adiposity and insulin sensitivity, or (in experimental models) endothelial nitric oxide depletion and augmentation of oxidative stress, have been observed with some of the newer antipsychotics (e.g. risperidone, olanzapine or clozapine; [31, 32]). These agents therefore seem to have the potential to cause complications that contribute to increased cardiovascular risk, necessitating careful benefit-to-risk assessment before use.

More advanced immune dysregulation as the major underlying cause of MHD, especially depression, might explain both the higher prevalence rates and the relative lack of benefit from antidepressants, particularly at more severe CVD stages. Thus, before antidepressant therapy is initiated, possible benefits should always be jointly evaluated by cardiologists and psychiatrists. Medications with anti-inflammatory properties (e.g. non-steroidal anti-inflammatories, cytokine inhibitors) have been shown to alleviate signs and symptoms of both CVD and depression [33]. In addition, recent clinical trials demonstrated that canakinumab and colchicine significantly reduce cardiovascular event rates in patients with coronary disease (reviewed by Nidorf et al. [34]). To what extent these drugs classes may also improve concurrent MHD requires further study.

The pathobiology of TTS is largely unknown, but suggested mechanisms include hypoconnectivity of central brain regions leading to altered limbic and autonomic processing in response to stress [40], and excess liberation of neurohormones and neuropeptides (e.g. neuropeptide Y) in response to stressful events [36]. The increased risk of TTS development in patients with pre-existing neuro-psychiatric illnesses might be due to both an exaggerated catecholaminergic response and increased sensitivity of the myocardium to catecholamines, all of which may exert direct cardiotoxic effects and induce microvascular dysfunction and myocardial stunning [36]. In line with the concept of sympathetic activation, elevated norepinephrine levels in the coronary sinus have been found in patients with TTS, suggesting increased myocardial catecholamine release [41]. In an in vitro induced pluripotent stem cell model of TTS, enhanced β‑adrenergic signalling and higher sensitivity to catecholamine-induced toxicity were identified as mechanisms associated with TTS [42]. In addition to myocardial dysfunction and injury, all these processes lead to dynamic ECG changes and moderate elevation of cardiac biomarkers in affected patients, and are referred to as “neurogenic stunning myocardium” (see Medina de Chazal et al. for more details; [36]).

Activation of the hypothalamic–pituitary–adrenal axis might link emotional stress and clinical TTS phenomena. In addition, pathways implicating female sex hormones could be important, given that 90% of TTS patients are women. Oestrogens may modulate catecholamine-induced vasoconstriction and upregulate endothelial nitric oxide (NO) synthase. Therefore, oestrogen supplementation may attenuate glucocorticoid and catecholamine responses to mental stress in perimenopausal women [43].

Diagnosing TTS is often challenging, and there is currently no non-invasive tool available to facilitate conclusive diagnosis of TTS. Instead, cardiac catheterization is the method of choice to exclude or confirm TTS. The InterTAK Diagnostic Score was developed by the International Takotsubo Registry to aid physicians in assessing the likelihood of TTS (Fig. 3; [44]), but diagnostic criteria continue to be disputed (see Ghadri et al. for more details; [35]). All components of this score can be obtained in the emergency department and do not require cardiac imaging. However, without invasive assessment, a definitive diagnosis is only possible by demonstrating reversibility of the condition.

The aims of treatment for TTS are decongestion and haemodynamic support, and interventions need to be tailored to individual presentation patterns [36]. For example, in hypertensive patients with left ventricular (LV) outflow tract obstruction (as may occur with apical ballooning), arterial vasodilators must be used with caution to avoid aggravation of this complication. Guideline-directed medical therapy (GDMT) is recommended. However, no data from larger randomized trials are available regarding the efficacy of renin-angiotensin-aldosterone system (RAAS) blockers and β‑blockers in TTS patients. As outlined by a recent review, pre-existing psychiatric comorbidity requires special attention [45]. The International Takotsubo Registry demonstrated that TTS patients had significantly more depression and anxiety than ACS patients in general, even after controlling for age and sex [35]. Rapid up-titration of certain psychotropic agents (e.g. serotonin-norepinephrine reuptake inhibitors [SNRI], lithium) that tend to increase endogenous catecholamines, or electroconvulsive therapy (which also induces a transient abrupt increase in catecholamine levels), may predispose to TTS development [45]. These observations further support the concept of a causal role of excess catecholamines in this syndrome.

Long-term management strategies for psychiatric comorbidities in TTS patients, including psychotropic measures, are poorly delineated. The use of various antidepressants known to increase catecholamine levels should be avoided because of their known association with incident TTS [35]. Selective serotonin inhibitors (SSRI) may be considered to treat depression and anxiety in TTS, but there is currently a lack of data in this area [45]. However, a retrospective study of 78 patients with TTS showed increased mortality and delayed recovery of LV function in those taking SSRI [46], an observation that may caution also against the use of this substance class. Overall, there is currently no evidence of therapeutic benefit from any psychotropic therapy in TTS in the literature, and treatments that could prevent recurrence of TTS are unknown.

The aetiology of PPCM is incompletely understood and likely multifactorial. Suggested, but yet unproven pathogenetic mechanisms are shown in Fig. 4. Prolactin (secreted from anterior pituitary gland) is cleaved by cathepsin D (activated by oxidative stress) into a 16-KDa fragment, which triggers endothelial dysfunction and apoptosis. Vascular damage leads to release of microRNA (mRNA)-146a and other mRNAs, which block important pathways thus inducing myocardial injury and metabolic insufficiency. Soluble fms-like tyrosine kinase receptor 1 (sFLT-1) produced by the placenta sequesters vascular endothelial growth factor (VGEF). Lack of VEGF leads to vascular and myocardial damage via several downstream pathways. Frequent occurrence of MHC is consistent with observations in severe CVD in general [5, 6]. However, in women with PPCM, impairment of the tryptophan metabolism has been reported, which increases the synthesis of quinolinic acid, with an associated reduction of serotonin synthesis, leading to reduced serotonin levels and increased production of the pro-inflammatory and MHD-promoting kynorenin, thus possibly providing a more specific explanation for the high incidence rates of MHC, especially depression, in women with PPCM [54]. This is compatible with the concept that (like certain infections) excessive psycho-physical stress may distort the interplay of the innate immune and central nervous systems, and implicates activation of toll-like receptors (e.g. TLR‑4, the transcription factor NF-kB, or the inflammasome NLRP3), as well as the secretion of interleukin‑1 beta, interleukin‑6 and other factors of the innate immune response, thus causing general symptoms of disease, but also adding to depression and anxiety [55]. Furthermore, patients with PPCM have higher levels of the depression-associated microRNA (miR)-30e [54]. Given that miR-30e impairs signalling pathways of the serotonin 1A receptor [56], it may also contribute to incident depression in women developing PPCM. Lastly, a connection between treatment with bromocriptine and MHD cannot be excluded [57].

Treatment of PPCM during pregnancy requires modifications to ensure foetal safety [47, 48]. After delivery, standard GDMT includes decongestive treatment, neurohormonal inhibitors, vasodilators and device therapies in selected patients in addition to specific pathophysiology-directed treatment with bromocriptine [58]. Improvement of HF symptoms as a result of these measures will likely also benefit the patients’ psychological situation. In view of the high prevalence of MHD in PPCM patients, psychiatric assessment and screening for post-traumatic stress disorder and other affective diseases should be part of the routine clinical evaluation. Clinical care strategies must include detailed needs analysis to facilitate personalized supportive strategies such as multimodal patient-centred management, which may also help prevent PPCM recurrence during later pregnancies. In the case of severe depression or anxiety disorders, targeted psychopharmacological and/or psychotherapeutic interventions are recommended, although systematic studies evaluating their efficacy in PPCM are lacking.

The increased risk of cognitive impairment associated with AF seems only partly explained by the generally accepted pathophysiological concept of a causal role of thromboembolic events, including both recurrent showers of microembolism and overt strokes [62]. Previous research described a decline in cognitive function in patients with incident AF even in the absence of clinical stroke, as well as a strong relationship between the duration of AF (AF burden) and the risk of dementia in younger patients [61]. Concordantly, the Atherosclerosis Risk in Communities (ARIC) study demonstrated that patients with high AF burden (persistent AF, 100%), but not those with lower AF burden (paroxysmal AF, 1–6%), had lower executive and verbal cognitive test scores compared with patients without AF [68]. Interestingly, the observed associations remained significant after adjustment for prevalent clinical stroke and persisted even after further adjustment for subclinical events, thus not conclusively explaining the relationship between a higher AF burden and lower cognitive function. Beyond cerebrovascular embolism, clinically unrecognized (silent) strokes, microbleeds, and/or other brain lesions might have contributed to these findings [68]. In a recent Korean nationwide cohort study in AF patients treated with catheter ablation or medical therapy, ablation was associated with decreased dementia risk [69]. Notably, this relationship was also evident after censoring for stroke and adjusting for clinical confounders [69]. Together, these observations provide evidence that AF may provoke cognitive decline, and that coexistence of AF and dementia is not a chance association. However, to what extent the rhythm itself is responsible for the cognitive impairment, and to what extent eliminating the arrhythmia will diminish the likelihood of dementia requires further investigation. Other factors, such as increased systemic inflammatory activity and oxidative stress accompanying AF may also contribute to worsening cognitive function. In summary, vascular dementia appears to be a heterogenous syndrome probably caused by a variety of mechanisms, which remain poorly understood beyond the role of cortical infarcts and cerebral small-vessel disease, which is undisputed [70, 71].

Further studies addressing underlying mechanisms are underway. Thus, in Swiss-AF the serum neurofilament light chain (sNfL), an acute phase marker of neuronal damage, was found to be associated with different cardiovascular risk factors (e.g. blood pressure and diabetes) as well as with vascular brain lesions and lower cognitive function [67]. In addition, baseline sNfL predicted brain atrophy after 2 years (unpublished data). At the 2‑year follow-up, new ischaemic brain lesions had occurred on MRI in 5.5% of patients; again, the vast majority of the lesions (85%) were clinically silent and had occurred in patients on oral anticoagulation therapy. New white matter lesions were found in 18%, and microbleeds in 11%. Overall, progressive cognitive decline was found in most tested dimensions if new brain lesions were apparent on repeat MRI (unpublished data). However, to what extent these observations are causally related to the presence of AF remains to be clarified.

Whether cerebral imaging, routine neurocognitive testing and routine analysis of markers such as sNfL should be used to screen for silent brain injury and whether implementation of such strategies into clinical routine care would improve prognosis is currently unknown. Recent findings suggest that effective oral anticoagulation may at least slow down cognitive decline and the development of dementia in patients with AF [72], and elimination of AF by ablation therapy seems to reduce the risk of dementia [69]. A more thorough understanding of the pathobiology of cognitive decline and dementia in patients with this common arrhythmia is required to promote development of efficacious preventive strategies.