Long COVID: post-acute sequelae of COVID-19 with a cardiovascular focus

Abstract Emerging as a new epidemic, long COVID or post-acute sequelae of coronavirus disease 2019 (COVID-19), a condition characterized by the persistence of COVID-19 symptoms beyond 3 months, is anticipated to substantially alter the lives of millions of people globally. Cardiopulmonary symptoms including chest pain, shortness of breath, fatigue, and autonomic manifestations such as postural orthostatic tachycardia are common and associated with significant disability, heightened anxiety, and public awareness. A range of cardiovascular (CV) abnormalities has been reported among patients beyond the acute phase and include myocardial inflammation, myocardial infarction, right ventricular dysfunction, and arrhythmias. Pathophysiological mechanisms for delayed complications are still poorly understood, with a dissociation seen between ongoing symptoms and objective measures of cardiopulmonary health. COVID-19 is anticipated to alter the long-term trajectory of many chronic cardiac diseases which are abundant in those at risk of severe disease. In this review, we discuss the definition of long COVID and its epidemiology, with an emphasis on cardiopulmonary symptoms. We further review the pathophysiological mechanisms underlying acute and chronic CV injury, the range of post-acute CV sequelae, and impact of COVID-19 on multiorgan health. We propose a possible model for referral of post-COVID-19 patients to cardiac services and discuss future directions including research priorities and clinical trials that are currently underway to evaluate the efficacy of treatment strategies for long COVID and associated CV sequelae.


Background
The crippling effect of coronavirus disease 2019 (COVID-19) on healthcare and economies globally has undoubtedly been one of the worst disasters experienced by humans in the last decades. Worldwide, survivors of COVID-19 now exceed hundreds of millions, 1 with some reporting incomplete recovery months beyond the acute illness, a condition commonly referred to as long COVID. Persistent symptoms of breathlessness, chest pain, fatigue, headaches, brain fog, and palpitations are a constant reminder of the devastation caused by this virus and the need to remain vigilant for any long-term damage. Now, more than ever before, management of cardiometabolic risk factors should become a priority for physicians, as their formidable power in intensifying COVID-19 illness severity has been convincingly documented. The long-term impact of COVID-19 on cardiovascular (CV) health and mortality is also emerging as a major global concern.
In this review, we discuss the definition of long COVID, epidemiology of cardiopulmonary manifestations in the context of long COVID, pathophysiological mechanisms for acute and chronic cardiac injury secondary to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, its management, and future directions.

Long COVID and epidemiology of cardiac symptoms
The term 'long COVID' was originally coined by a patient 2,3 and asserts the notion that suffering does not stop with resolution of acute infection. While there is no universally accepted definition, in December 2020 the United Kingdom (UK) National Institute for Health and Care Excellence guidelines 4 defined long COVID as persistence of symptoms beyond 4 weeks of SARS-CoV-2 infection. This term comprises two phases: ongoing symptomatic phase (4-12 weeks) and post-COVID-19 syndrome (.12 weeks) based on the duration of symptoms. More recently, the World Health Organization provided a case definition for post-COVID-19 condition, 5 a term used to refer to persistence of symptoms beyond 3 months of SARS-CoV-2 infection, lasting for at least 2 months and not explained by any other illness. Other terms used to describe long COVID include post-acute COVID-19 syndrome, 6 post-acute sequelae of COVID-19, 7 and long-haul COVID. 8 Long COVID is a vacillating disease, 9 characterized by a diverse range of symptoms spanning multiple organ systems, as depicted in Figure 1 and Graphical Abstract, and commonly includes fatigue, breathlessness, post-exertional malaise (PEM), brain fog, headaches, nausea, vomiting, anxiety, depression, skin rash, joint pain, and palpitations. Patient advocacy groups 6 (e.g. long COVID SOS, COVID Advocacy Exchange, the National Patient Advocate Foundation COVID Care Resource Center, long-haul COVID fighters, Body Politic COVID-19 Support Group) have enhanced our understanding of this disease by drawing our attention to its multifaceted nature. Several experts 10,11 have noted its marked similarities with other post-viral symptoms (e.g. Epstein-Barr, 12 human herpesvirus, 13 influenza, SARS, 14 and Ebola viruses 15 ), although few options exist for the management of such syndromes.
The reported prevalence of long COVID has varied across and within many countries: UK 1.6-71%, [16][17][18][19] Germany 35-77%, 20,21 China 49-76%, 22,23 Africa 68%, 24 India 22%, 25,26 Bangladesh 16-46%, 27,28 Denmark 1%, 29 Italy 5-51%, 30,31 USA 16-53%, 32,33 Norway 61%. 34 Studies assessing hospitalized patients have typically reported higher prevalence estimates (e.g. 76% in Huang et al., 22 71% in Evans et al. 19 ) when compared with community studies (e.g. Sudre et al. 16 ), reflecting the complex relationship between severity of acute illness, higher burden of co-morbidities, and persistent symptoms. Differences in the study population may explain the vast disparity in prevalence estimates across the various studies. The timing of assessment also appears to be important as symptom frequency can diminish over time from the infection. Hossain et al. 28 reported a reduction in the burden of long COVID symptoms which affected 21.2% of their cohort at 4 weeks and 16.5% by 12 weeks post-COVID diagnosis. A similar temporal improvement in symptom burden was also observed by Wu et al. 35 and Cassar et al. 36 Varying definitions of long COVID may also affect the relative frequencies. Mahmud et al. 27 defined long COVID as the persistence of symptoms beyond 2 weeks (i.e. time taken for viral clearance) and reported a symptom prevalence of 46%. In contrast, more conservative definitions such as the one used by the UK Office of National Statistics (requiring the presence of functional limitation and exclusion of symptoms explained by comorbidities) have resulted in lower prevalence estimates. The study design is also likely to be relevant as retrospective reporting from electronic healthcare records suffers from ascertainment bias, while prospective studies with comprehensive assessments 37 are likely to attract patients with a high burden of symptoms seeking an explanation. In addition to these, disparities in vaccinations, SARS-CoV-2 variants, co-morbidities, study sample size, and use of varying non-COVID control groups appear to drive heterogeneity in prevalence estimates, as illustrated in Figure 2.
Contrary to the variability seen in reported disease prevalence, risk factors for long COVID tend to be fairly consistent, with female sex, escalating age, obesity, asthma, poor general health, poor prepandemic mental health, poor sociodemographic factors emerging as important determinants across several studies. [16][17][18][38][39][40] In particular, the impact of nationwide lockdowns, remote working, and limited physical activity on pre-existing trends of an increasingly obese population with poor dietary intake and physical activity patterns is noteworthy. [41][42][43] According to national statistics data from the UK (from 2019), 44 among adults 16 and over, a staggering 68% of men and 60% of women were either overweight or obese. Obesity increased across all age groups up to 75 years old. In a separate report by the national child measurement programme from the UK, one in three children leaving primary school were noted to be overweight or obese and one in five obese in 2019. 45 A recent update from the American Heart Association (AHA) 46 on stroke and CV disease has also highlighted the high prevalence of obesity, metabolic syndrome, poor dietary habits, and physical inactivity among children and adults from the US. It is now well established that obesity and other cardiometabolic risk factors commonly promote inflammation and endothelial dysfunction, 47,48 which may lower the cardiometabolic reserve and threshold for exertional symptoms. Consistent with this, numerous population and prospective cohort studies have documented an independent link between obesity and long COVID. 16,19,38 Thompson et al.,38 in a prospective study of 6907 patients (mean age 19-63 years), reported that being overweight or obese was associated with a 25% higher likelihood of long COVID than not belonging to this category. Similarly, Sudre et al. 16 reported that patients with prolonged symptoms were more likely to be obese than those without.
Cardiopulmonary symptoms including chest pain, dyspnoea, fatigue, palpitations, and cough are common among long haulers. In one UK study, 16 13.3% of 4182 symptom app users (predominantly community patients) experienced at least one persistent symptom beyond 4 weeks of infection, of which half were thought to be cardiac in origin. In December 2020, the UK ONS 49 provided similar estimates of long COVID prevalence, though a recent Figure 1 Long COVID is characterized by a diverse range of symptoms and signs spanning multiple organ systems including the respiratory system, neurological system, cardiovascular system, gastrointestinal system, dermatological system, endocrine/genitourinary systems, and skeletal muscle/joints as illustrated here. PTSD, post-traumatic stress disorder.
Long COVID/post-acute sequelae of  analysis with a stricter case definition (symptoms explained by comorbidities were not counted) by the ONS 18,50 and other groups 29 suggests lower prevalence estimates (1.2-1.5%). Patient-led or survey-based research from long COVID support groups 37,51 has also provided insights into the longitudinal trajectory of persistent symptoms. In an international online survey study undertaken by Davis et al. 37 of 3762 patients, cardiac symptoms including chest pain ( 53%), palpitations ( 68%), fainting ( 13%) were observed in up to 86% of patients by 7 months from infection. Postural orthostatic tachycardia syndrome (POTS), characterized by an increase in heart rate of at least 30 b.p.m. from supine to standing position, was noted in 31% of patients. 52 Ziauddeen et al. 51 similarly studied the prevalence of long COVID among 2550 patients using a social media survey. Cardiopulmonary symptoms were reported by 89% of participants in their study.
There are now several prospective follow-up studies of hospitalized patients. One of the earliest published reports of long COVID came from an Italian study of hospitalized patients by Carfi et al., 53 demonstrating a particularly high burden of cardiopulmonary symptoms (.43%). In a subsequent follow-up study of 1733 hospitalized patients from Wuhan, China, Huang et al. 22 observed that at 6 months post-infection, 63% of patients reported fatigue, 26% breathlessness, and 5-9% experienced chest pain and palpitations. By 12 months, 23 investigators of the same study showed that symptoms of breathlessness (30%) and chest pain (7%) were slightly more common, while fatigue had improved (20%). Evans et al. 19 from the UK also undertook a follow-up study of 1077 hospitalized patients. At a median of 5 months postdischarge, 48% reported persistent fatigue, 41% dyspnoea, and 21-28% chest pain and palpitations.
Long COVID has been proposed to be a form of chronic fatigue syndrome (CFS)/myalgic encephalitis. 52 While there are marked similarities between the two, subtle yet important differences exist. Central to the diagnosis of CFS is the observation that patients experience PEM, 54,55 defined as fatigue following even minor physical or mental exertion. For a diagnosis of CFS, symptoms should last for a minimum of 6 months and occur at least 50% of the time. Current definitions of long COVID do not obligate the presence of Figure 2 Factors that contribute to variability in prevalence estimates of long COVID. Prevalence estimates of long COVID are highly variable across studies due to a number of factors that introduce bias. These include differences in cohort characteristics, age, and sex of subjects enrolled, timing of assessment, sociodemographic factors, vaccines and variants, pre-existing health problems, sample size, study design, and variability in questionnaires or tools used. HCW, healthcare workers; m, months; PCR, polymerase chain reaction; wks, weeks. PEM, with symptom duration yet to be defined. A further point to note is that, while the evidence for exercise rehabilitation in CFS patients is mixed, early positive data supporting tailored rehabilitation in previously hospitalized patients with long COVID are emerging with some improvement seen in exercise capacity and cognition at 4 months from discharge. 56 Finally, breathlessness, one of the most common long COVID symptom, is not essential for the diagnosis of CFS and may be due to distinct mechanisms. Nonetheless, the similarities between CFS and long COVID are striking, with both syndromes having links to psychological and neurostructural/metabolic alterations 57-59 (e.g. grey matter volume reduction in the limbic cortex in post-COVID patients 60 ), highlighting the potential benefits of neuroprotective interventions in such patients.

Possible mechanisms underlying cardiac manifestations
Given the abundance of cardiac symptoms among patients with long COVID, a deeper discussion of potential mechanisms underlying cardiac injury is warranted.

Acute phase
The role of angiotensin-converting enzyme 2 receptors in SARS-CoV-2 involvement of the heart is now well established. Several mechanisms have been proposed to contribute to myocardial injury including direct cytotoxic injury, 1 dysregulation of renin-angiotensin-aldosterone system, 3 endotheliitis and thromboinflammation, and 4 dysregulated immune response with cytokine release. 6

Autopsy studies
The pattern of myocardial injury following SARS-CoV-2 infection derived from autopsy studies suffers from referral bias but has provided initial pathophysiological clues. In an early autopsy series of 80 consecutive SARS-CoV-2 PCR positive cases, 61 only four patients (5%) had suspected cardiac injury. Two patients had co-morbid conditions and died of sudden cardiac death. One had acute myocardial infarction and another exhibited right ventricular lymphocytic infiltrates. These early results suggested that extensive myocardial injury as a primary cause of death may be infrequent.
In a subsequent multicentre autopsy study, Basso et al. 62 characterized the hearts in 21 selected autopsies. Myocarditis (defined as lymphocytic infiltration and myocyte necrosis) was evident in 14% of cases, interstitial macrophage infiltration in 86%, and pericarditis and right ventricular injury in 19%, respectively. Halushka and Vander Heide 63 performed a review of 22 publications describing the autopsy results of 277 patients. Lymphocytic myocarditis was reported in 7.2%; however, only 1.4% met the well-established histological criteria 64 for myocarditis, suggesting that true myocarditis was rare. [65][66][67] In another study, Lindner et al. 68 demonstrated the presence of SARS-CoV-2 viral particles in the hearts of 24/39 (59%) consecutive autopsies; the viral load was clinically relevant in 16/39 (41%). Of note, viral particles were not isolated within cardiomyocytes, but rather in interstitial cells including pericytes and macrophages. The high viral load in some cases was also not associated with inflammation, consistent with the low prevalence of myocarditis on autopsy studies. 67 Microvascular injury due to SARS-CoV-2 Cardiac troponin levels are frequently elevated in COVID-19 patients, 69 indicating myocardial injury and/or ischaemia. The work of Bois et al. 70 appears to support the concept of microthrombi occurring in association with COVID-19. In a small series of 15 individuals, the authors observed that post-mortem fibrin microthrombi were more common (80%) than acute ischaemic injury (13%) and myocarditis (33%) suggesting a role for thrombosis in accentuating myocardial injury.
Fox and Vander Heide 65 have conceptualized the array of pathophysiological mechanisms underlying myocardial injury. The authors proposed that hypoxia and pulmonary microvascular damage may lead to right heart stress and myocyte necrosis. The latter may be further augmented through localized microvascular effects, endotheliitis, 71 associated microthrombi, and altered renin-angiotensin homeostasis. 72 Elevated cytokines 73,74 [e.g. interleukins (IL)-1, IL-16, IL-17, IL-22, interferon (IFN)-γ, tumour necrosis factor (TNF)-α] could also contribute to myocardial injury by inducing endothelial dysfunction, activation of platelets, recruitment of neutrophils, and eventually triggering a hypercoagulable state. In this framework, viral myocarditis would play an infrequent role in SARS-CoV-2 infection.

Post-acute and chronic phase
Mechanisms for enduring cardiac damage post-acute illness are still poorly understood. One possible explanation is a chronic inflammatory response evoked by persistent viral reservoirs in the heart following the acute infection, 75 which may, in turn, be exacerbated by obesity-related inflammatory signalling driven in part by perivascular adipose tissue via the release of adipokines such as monocyte chemoattractant protein-1 and Regulated upon Activation, Normal T Cell Expressed, and Presumably Secreted, chemokines that aggravate endothelial dysfunction via endothelial nitric oxide synthetase uncoupling and reactive oxygen species production. 76 An unintended consequence of such processes would be insidious tissue damage, followed by chronic myocardial fibrosis leading to impaired ventricular compliance, impaired myocardial perfusion, increased myocardial stiffness, reduced contractility and potential arrhythmias.
A second mechanism for delayed damage is an autoimmune response to cardiac antigens through molecular mimicry. 77 High-throughput proteome analysis by Wang and others [78][79][80] has identified a range of autoantibodies to humoral and tissue antigens in patients with severe COVID-19. Autoantibodies to cholinergic and adrenergic receptors have also been detected in individuals with CFS. 81,82 Recently, a number of longitudinal cytokine profiling and proteomic studies 83,84 have revealed an increased expression of prothrombotic factors (e.g. factor VIII, prothrombin, plasminogen activator inhibitor-1) beyond the acute infection. This is in keeping with the burgeoning reports of delayed embolic complications. [85][86][87] The high prevalence of pulmonary vascular thrombosis (5-30%), 87,88 particularly in hospitalized patients, is also expected to heighten the future risk of chronic Long COVID/post-acute sequelae of COVID-19 thrombo-embolic pulmonary hypertension. 89 Endothelial dysfunction 90 and its complications may also develop in patients, with evidence of persistent impairment detected in younger individuals 3-4 weeks after SARS-CoV-2 infection. 91 The Graphical Abstract summarizes the relevant pathophysiological mechanisms for acute and chronic cardiac injury secondary to SARS-CoV-2 infection and potential long-term consequences. It is worth noting that many of these complications mirror those encountered by survivors of other epidemics caused by SARS, Middle East respiratory syndrome (MERS), H1N1A, underscoring the need to recognize the impact of respiratory viral infections on CV health as previously outlined by Xiong et al. 92 in an earlier review.

Post-acute COVID-19 cardiovascular sequelae
The high mortality and poor outcomes associated with myocardial injury 93-97 during acute COVID-19 infection have galvanized an interest among research communities to characterize the long-term CV effects of SARS-CoV-2 infection. Insights from both prospective and retrospective studies continue to shape our understanding of its long-term effects. While retrospective studies rely on electronic medical health records and labelled datasets, 7,34,98 prospective studies 19,22 have innovatively turned to remote (e.g. telemedicine, symptom apps) assessments and face-face reviews.

Retrospective cohort studies
There is now compelling evidence from large retrospective cohort studies that speak to the rising cases of new cardiac diagnoses. In a study of 73 435 (median age 61 years, 88% men) non-hospitalized patients using the US Department of Veterans Affairs health services, Al-Aly et al. 7 demonstrated a high risk of death and incident CV and metabolic diseases associated with COVID-19 beyond 30 days of infection. A UK-based study of 47 780 hospitalized COVID-19 patients (mean age 65 years, 55% men) demonstrated that a diagnosis of COVID-19 was linked to a three-fold increased risk of major adverse CV events up to 4 months from diagnosis (vs. non-hospitalized controls). In this study, Ayoubkhani et al. 98 further noted that the increased risk was not confined to the older age group and was more pronounced in non-White patients. Daugherty et al., 99 in a related paper, compared the incidence of new cardiometabolic diagnoses in post-COVID-19 patients with two important controls groups-non-COVID-19 controls (from 2019 to 2020) and those recovering from lower respiratory tract infection (LRTI). In their study, COVID-19 was associated with a nearly two-fold increased risk of incident CV diagnoses. However, when comparisons were made with LRTI controls, the excess risk of cardiomyopathy was no longer significant. These findings are in line with another study that used primary healthcare data in the UK (OpenSAFELY platform). Tazare et al. 100 revealed that the excess risk of major adverse CV events among previously hospitalized COVID-19 patients was similar to patients admitted with a diagnosis of pneumonia, although the risk of developing type 2 diabetes was higher after COVID-19.

Prospective studies
In the post-acute period, cardiac abnormalities have been reported in several prospective observational studies.  A summary of selected studies (n . 50) employing three widely used investigative tools-echocardiography, cardiac magnetic resonance (CMR), and cardiopulmonary exercise test-is provided in Table 1 101,102,109,112,[114][115][116][120][121][122]124,125,128,130,131,133,135 and highlights the vast heterogeneity in the prevalence of abnormalities.
Numerous studies have evaluated the role of 12-lead electrocardiogram (ECG) in screening patients for post-acute cardiac manifestations, 36,103,117,124,136 although commonly pre-COVID-19 control ECGs are missing. Dynamic ECG changes (e.g. depolarization, repolarization abnormalities, arrhythmias) 137,138 while frequent during acute illness, tend to resolve in the majority of hospitalized patient by 6 months post-acute COVID-19 and are often comparable to risk-factor matched controls. 36,103,117,136 Nevertheless, sinus arrhythmia is frequent in the post-acute phase and manifests as transient or sustained periods of sinus tachycardia or bradycardia. [103][104][105] In a study of 234 patients, Radin et al. observed that 13.7% of patients wearing a Fitbit displayed persistently elevated heart rate (.5 beats above the pre-COVID baseline resting heart rate) up to 133 days postinfection. 139 Currently, there are no published studies on the role of prolonged ECG monitoring (Holter) in post-COVID-19 management. Prior studies of post-influenza patients have demonstrated a high burden of atrial 140 and ventricular arrhythmias, which are known to correlate with inflammatory markers. [141][142][143][144][145] These findings imply that a subset of COVID-19 patients (e.g. with ongoing inflammation) may potentially benefit from ECG monitoring in the long term.
Both transthoracic echocardiography and CMR are cornerstones in the diagnosis of acute and chronic cardiac pathology. 146,147 Cardiac abnormalities commonly reported on follow-up imaging include myopericarditis, right ventricular dysfunction, and ischaemia/infarction.
Myopericarditis may be suspected clinically based on the presence of one clinical (pericarditic chest pain, heart failure or progression of heart failure symptoms, palpitations, syncope, newonset arrhythmia) and one diagnostic criteria (ECG abnormalities, troponin elevation, wall motion abnormalities on echocardiogram, CMR abnormalities) as per the 2013 European Society of Cardiology (ESC) Position statement 148 and 2020 AHA scientific statement. 149,150 While endomyocardial biopsy (EMB) is the gold standard investigation for histological evaluation of suspected fulminant cases, [148][149][150] CMR remains the best alternative for noninvasive evaluation of stable cases 149,151 and provides information on several pathological processes including myocardial oedema, hyperaemia, necrosis, and fibrosis by exploiting alterations in fundamental magnetic properties of the tissue (T1 and T2 relaxation). In 2018, the Journal of American College of Cardiology scientific expert panel provided recommendations for updated CMR criteria to improve detection of active myocarditis. 151 This required an increase in at least one T1-based method including T1 mapping (sensitive to hyperaemia, fibrosis, necrosis, and oedema) and one T2-based method including T2 mapping (sensitive to oedema). 151 Although the incorporation of quantitative tissue mapping has augmented the sensitivity of CMR for myocarditis, two limitations  exist with such an approach. The first is that CMR diagnostic performance may be impacted by co-morbidities. 152,153 The second is that readouts of T1 and T2 values of the myocardium are poorly standardized. 154 As a result, combining quantitative CMR data from multiple centres is problematic. In June 2020, a follow-up CMR study of 100 patients (67% nonhospitalized) reported an alarmingly high rate (60%) of persistent myocardial inflammation at 71 days post-infection. 122 At least 22% of patients in this study were found to be co-morbid. Other studies have also reported a high prevalence of CMR abnormalities, though comparator groups were typically healthy making it challenging to rule out confounding effects of co-morbidities such as hypertension and diabetes. In contrast to these early studies, a subsequent study 128 of healthcare workers, which enrolled comorbidity matched controls with mild infection, reported a lower prevalence of CMR abnormalities, with no significant difference in 6-month CMR tissue abnormalities between seropositive and seronegative healthcare workers. Similarly, a study of 1285 UK Biobank participants 155 with pre-and post-SARS-CoV-2 infection imaging revealed no link between prior infection and longitudinal changes in cardiac or aortic phenotypes before and after adjusting for potential confounders including co-morbidities.
There are now abundant studies evaluating the burden of myocarditis among athletes in view of associated risks of sudden cardiac death. The majority was undertaken within 1-2 months of infection 105,126,127,129 and the prevalence of myocarditis on objective testing was found to be generally low (0-3%). Studies evaluating persistent symptom burden among athletes beyond 8 weeks of infection are lacking, though it is likely, given the lower risk profile of athletes (e.g. less likely to be obese and co-morbid), that ongoing symptoms are infrequent in this population.
The role of CMR in elucidating a cause for elevated troponin following acute COVID-19 is less contentious. Several prior studies [156][157][158] have confirmed its incremental value in clarifying a diagnosis when the aetiology of troponin elevation is unclear. In a study of 148 troponin-positive patients, Kotecha et al. 101 reported that 26% had myocarditis, 22% had inducible ischaemia/infarction, and 6% had evidence of both.
Echocardiography is vital in the early diagnosis of cardiac pathology in COVID-19 infection (suspected myocarditis, Takotsubo syndrome, myocardial infarction, pericardial effusion, etc.), particularly where haemodynamic stability is uncertain. 69,159 Right ventricular dilatation and dysfunction are the most common echocardiographic abnormalities with prognostic significance. 125,160,161 When the acute infection abates, 161 right ventricular abnormalities improve in most patients. 36,125 Left ventricular systolic dysfunction is comparatively less frequent. 101,125 Follow-up echo and CMR studies have confirmed that even among those with severe acute infection 125 or elevated troponin, 101 systolic impairment is rare, affecting up to 9-11% of patients. Concordant with this, in the UK-wide national follow-up study (PHOSP-COVID), levels of N-terminal pro B-type natriuretic peptide (NT-proBNP) were abnormal in only 7% of patients at 5 months post-hospitalization. 19 Unlike systolic dysfunction, abnormalities in diastology may be common (up to 60% of hospitalized patients). 109 The lack of pre-COVID imaging, however, makes it challenging to disaggregate what is cause and effect in patients.
Computed tomography (CT) angiography (CTA) has garnered considerable attention for its ability to detect pulmonary emboli, 162-164 epicardial coronary stenoses, or vascular pathology (e.g. mural thrombus or vasculitis) related to acute SARS-CoV-2 infection. Perivascular fat attenuation index, 165,166 a biomarker of vascular inflammation on CTA, has shown promise for prognostic risk stratification. In an early study by Kotanidis et al., 167 a new radiotranscriptomic signature of vascular inflammation demonstrated an association between SARS-CoV-2 variant B1.1.7, vascular inflammation, and increased mortality risk. Delayed arterial and venous thrombo-embolic complications [168][169][170] have also been reported in the post-acute period. Dual-energy CTA, in a study of 55 patients at 3 months post-infection, detected both proximal arterial thrombosis (5.4%) and distal microangiopathy (65.5%) in a significant proportion of symptomatic patients. 171 Cardiopulmonary exercise testing has shed light on the relevant pathophysiological brakes applied by COVID-19 on exercise capacity. 36,102,[106][107][108]115,[118][119][120] Several studies have demonstrated a reduction in peak oxygen consumption post-acute COVID-19. 36,102,106,115,120,172 The predominant mechanism for this finding seems to be muscular deficiencies (impairment in oxygen extraction 173 ), manifesting primarily as submaximal exercise tests or an early anaerobic threshold. Generalized muscle wasting or sarcopenia is also common. 174,175 Physical inactivity, cytokine storm, poor nutrition, intensive care therapy, mechanical ventilation, and drugs (e.g. dexamethasone) have all been implicated. The ratio of exercise minute ventilation coupled with carbon dioxide output (VE/VCO 2 ), a marker of ventilatory efficiency, may also be abnormal. 36,102,120 However, breathing reserve is relatively preserved, arguing against pulmonary factors and supporting hyperventilation or dysfunctional breathing as a potential cause. 172 Heart rate recovery provides a surrogate measure of autonomic health in patients. Following COVID-19, delayed heart rate recovery has been noted in some patients, 36,176 though the majority recovers spontaneously over time. 36 There are now numerous reports [177][178][179][180][181] and cohort studies 38 where POTS has been suspected among patients. In a retrospective study of 20 patients 182 referred to a dysautonomia clinic, orthostatic instability on tilt table test or 10 min stand test was observed in 75% of patients. Postural orthostatic tachycardia syndrome was the most common diagnosis in this study, followed by neurocardiogenic syncope (15%) and orthostatic hypotension (10%).
A table summarizing all the relevant cardiac investigations, their advantages, and role in post-COVID management can be found in the Supplementary material online, Table S1.

Cardiovascular injury in the context of multiorgan involvement
Understanding cardiac involvement in the context of multisystem health can provide clues into mechanisms of ongoing injury through recognition of unique patterns of tissue damage (e.g. inflammatory changes or embolic manifestations). In an early study of 58 post-Long COVID/post-acute sequelae of COVID-19 hospitalized COVID-19 patients and 30 matched controls, Raman et al. 120 undertook multiorgan magnetic resonance imaging (MRI) and reported tissue abnormalities involving the lungs (60%), heart (26%), liver (10%), kidneys (29%), and brain (11%) in patients. Magnetic resonance imaging abnormalities in almost every organ correlated with inflammatory markers, suggesting that chronic inflammation could impede recovery. Following on from this work, the PHOSP-COVID study also demonstrated that failure to recover from multiorgan symptoms was associated with markers of persistent inflammation. 19 In another study of 201 patients, Dennis et al. 121 evaluated the prevalence of multiorgan damage among predominantly non-hospitalized patients and noted that symptoms of long COVID clustered among those with multiorgan injury on MRI.
Persistent endothelial dysfunction, 47,183,184 microvascular dysfunction, 101 and prothrombotic tendencies 185 may also contribute to multiorgan dysfunction. 186 Selected studies of advanced imaging modalities including positron emission tomography, CT, and MRI have noted perfusion deficits in the heart and lungs 101,187,188 of COVID-19 survivors at 40-60 days from infection. In one study of hospitalized patients, multiorgan MRI demonstrated evidence of small vessel disease (9.3%) and ischaemic changes (3.7%) in the brain (9.3%) and 1.9% had myocardial infarction 120 2-3 months postinfection. Another study of 104 hospitalized 189 patients observed that inducible myocardial perfusion defects were common among patients with moderate to severe disease but did not differ in burden compared with co-morbidity and risk-factor matched controls. Several studies are currently underway to characterize the burden of vascular and thrombotic complications 190,191 and to examine the potential benefits of prolonged antithrombotic (extended thromboprophylaxis) and vascular protective therapies (e.g. statins, risk-factor management) in post-acute COVID-19 patients as indicated in Table 2 and Supplementary material online, Table S2.

Chronic cardiovascular disease and impact of COVID-19
Up to a third of patients hospitalized with COVID-19 have a history of chronic CV diseases. 192,193 Presence of concurrent cardiac disease is typically associated with higher in-hospital mortality, thrombo-embolic risk, and septic shock rates. 97,194 Even in the post-acute period, patients with a history of heart failure are at a two-to four-fold risk of decompensation and mortality. 195,196 Increased rates of heart failure exacerbation may present even beyond 30 days after SARS-CoV2 infection. 98 One reason for the rising epidemic of post-discharge heart failure events is the withdrawal of guideline-directed medical therapy during acute illness. 197 Investigators of the TRED-HF study 198 have previously demonstrated the negative impact of heart failure therapy withdrawal in recovered dilated cardiomyopathy patients, resulting in relapse and poor outcomes. The successful resumption and optimization of heart failure therapies 199 may therefore be important in halting heart failure readmissions post-acute COVID-19.
The shared cardiometabolic profile of COVID-19 and cardiac diseases implies that COVID-19 may play a role in destabilizing subclinical diseases (e.g. coronary artery disease and heart failure). This would explain the high prevalence of type 2 myocardial infarction 200 in those with severe COVID-19 and also the rising incidence of 'new' CV diagnoses. 7,98,99 Other mechanisms that may contribute include dysregulation of the renin-angiotensin-aldosterone system, 6 endothelial dysfunction, 186,201,202 renal injury, 203 and steroid use. 204,205 Proposed model for investigation and management of cardiovascular sequelae and long COVID Although the true burden of CV pathology post-acute COVID-19 remains elusive, the prevalence of cardiac symptoms in this phase appears high. There is a strong need for evidence in support of cost-effective strategies to exclude significant CV pathology. An approach considered reasonable by some experts [206][207][208] involves screening of high-risk individuals for ongoing cardiac involvement including those with abnormal cardiac investigations during the acute phase, new CV diagnosis post-COVID-19, and athletes. Figure 3 shows a possible pragmatic algorithm to guide physicians on the indication for cardiology follow-up and management approaches. Screening of high-risk individuals could comprise of a thorough history, clinical examination, blood test panel (C-reactive protein, troponin, B-type natriuretic peptide/ NT-proBNP, glycated haemoglobin, lipids), ECG, and transthoracic echocardiography at least 8-12 weeks from infection. For patients with clinically significant abnormalities after the screening, additional testing is recommended. Non-invasive tests such as CMR, stress single positron emission computed tomography, Holter, coronary CTA can be considered following screening investigations; invasive coronary angiography or EMB may be indicated for high-risk individuals. Referral to specialist clinics (e.g. POTS, arrhythmia clinic, psychology support) should be considered where relevant. Patients with chronic CV diseases presenting for routine follow-up should be asked about their history of COVID-19 infection and vaccination status. A brief assessment of mental, physical, and cognitive health may be required for selected patients who report ongoing symptoms as this could facilitate early referral to appropriate services (rehabilitation, 209 physiotherapy, 210 psychology, 211 occupational therapy, and social and welfare support) ( Figure 3) and alleviate patient burden.
With regard to return-to-play guidance for athletes, numerous recommendations have been put forward by consensus societies. [212][213][214] Although earlier guidelines adopted a conservative approach, recent studies of college and professional athletes have led to a revision 126,129 in recommendations. Graded resumption to exercise and sports is now considered reasonable for mild infections, whereas exercise restriction for 3 months is still recommended for individuals with suspected myocarditis as per the 2019 position statement from the Sport Cardiology Section of the European Association of Preventive Cardiology. 215 The management of COVID-19-related chronic myocarditis is a subject of considerable debate. In individuals with complicated    or tachyarrhythmias), the ESC 148 and AHA 149 recommend EMB for clarification of myocarditis subtype to guide specific treatment options (e.g. immunomodulatory therapy vs. antivirals). 150 Currently, there is no COVID-19 specific guidance on this, though several studies are underway to evaluate the most effective management strategy. The efficacy of oral non-steroidal antiinflammatory drugs and/or colchicine is also being evaluated for COVID-19-associated pericarditis. 216 For the management of post-COVID-19 acute coronary syndromes, patients are typically treated in accordance with the ESC 217 and AHA 218 guidelines released in 2020 and 2014, respectively. Similarly, heart failure management revolves around optimal utilization of contemporaneous therapies as per guidelines. 219,220 There are currently no published trials on the efficacy of prolonged thromboprophylaxis post-acute COVID-19; however, numerous intervention trials (e.g. HEAL-COVID, 221 STIMULATE ICP) 216 are currently ongoing to address this gap.
After exclusion of significant CV and other organ pathology, the management of long COVID tends to be largely supportive. 222 Given the strong association between obesity and long COVID, measures to reduce weight through caloric restriction, diet, tailored graded exercise, stress reduction, and good sleep hygiene could be beneficial in the long run, 223 with growing evidence indicating its favourable effects on systemic inflammation, 224 vascular dysfunction, 225 and metabolic syndrome. 226 Additionally, a pragmatic approach that is holistic and targeted at alleviating symptoms may also be required. Non-pharmacological approaches including pulmonary rehabilitation, 227 breathing exercise, 228,229 and alternative therapies 230 (e.g. singing therapy, 231 acupuncture, body rotation, and stretching) have also been suggested to help breathlessness symptoms. Those returning to work may benefit from phased return, allowing individuals with incomplete mental and physical recovery to gradually resume employment. 232 Given that psychosocial factors are a major determinant of incomplete recovery, 38,233 early referral for mental health assessment/cognitive behavioural therapy may benefit some patients.
Postural orthostatic tachycardia syndrome and symptoms of dysautonomia can be debilitating for patients. [178][179][180][181] The management of POTS centres around accurate diagnosis following specialist assessment, correction of reversible causes (dehydration, heat), optimization of chronic disease management, and patient education. In some patients with ongoing palpitations, beta-blockers can be helpful in treating symptoms. 234 Graded exercise programmes 235 encouraging patients to adopt an upright posture may attenuate postural symptoms after prolonged bed rest. Compression pantyhose style stockings with 30-40 mmHg counter pressure may help symptoms of orthostatic hypotension through reduced peripheral venous pooling. 236 In the event that symptoms persist despite compliance with the aforementioned approaches, pharmacological therapies (e.g. ivabradine, fludrocortisone, midodrine, clonidine, and methyldopa 237 ) may be considered.

Impact of acute therapies
The vast majority of persistent CV abnormalities following COVID-19 is due to tissue injury sustained during the acute illness. The impact of acute therapies on long-term CV health deserves further investigation. Currently, anti-inflammatory drugs such as dexamethasone 238 and tocilizumab 239 or antivirals such as remdesivir 240 have been identified as key weapons in the therapeutic armamentarium against severe COVID-19. However, the extent to which they affect cardiopulmonary recovery in the long term is still unclear and data regarding cardiac injury rates are not yet widely available. Whether or not ongoing inflammation in long COVID may reflect a rebound phenomenon in dexamethasone-or tocilizumab-treated patients also warrants further study. The complicated role of anticoagulation in patients deserves some consideration. In the acute phase, mounting evidence confirms a lack of benefit of aspirin in reducing mortality among hospitalized 241 and non-hospitalized outpatients. 242 Data in support of therapeutic dose anticoagulation are mixed with the severity of illness (non-critical hospitalized patients benefitting most) being a critical determinant of treatment success. 243 The multiplatform adaptive randomized controlled clinical trial, 244 which combined data from ACTIV-4a, REMAP-CAP, and ATTACC studies, reported an improved survival until hospital discharge and organ support-free days with therapeutic dose heparin in moderately ill patients but not during critical illness. In contrast, other studies (ACTION, 245 INPIRATION, 246 and RAPID trials) 247 reported no difference in primary outcome measures among patients receiving therapeutic vs. prophylactic dose anticoagulation. Investigators of the ACTIV-4B 242 study also reported no improvement in 45-day survival among non-hospitalized patients receiving aspirin, low dose, and high dose apixaban vs. placebo. Further research is therefore needed to better understand the long-term benefits of anticoagulation in patients.

Role of vaccination-risks vs. benefits
The most effective way of preventing serious complications from SARS-CoV-2 infection is through vaccination. [248][249][250][251][252][253][254] Previous experience with influenza vaccine has taught us to expect a favourable relationship between vaccination and CV outcomes. 255,256 There are at least eight major SARS-CoV-2 vaccines available globally, with excellent efficacy. Early data from a patient-led observational study 257 has hinted at the possibility of long COVID symptoms being alleviated through vaccination. Of 900 people with long COVID, 56.7% of those vaccinated saw an overall improvement, 18.7% a deterioration, and 24.6% were unchanged post-vaccination. In another survey study (COVID symptom app study), 258,259 the odds of experiencing symptoms more than 28 days post-vaccination was halved by two vaccinations (n = 906). Some experts posit that an accelerated viral clearance and a muted chronic inflammatory response could explain symptom reduction following vaccination. 260 While there is little doubt that early inoculation confers the greatest protection against severe COVID-19, rare cases of vaccine-induced adverse effects have led to a rise in vaccine hesitancy. 261 In particular, two widely available vaccine strategies including mRNA vaccines 250,262 and vector-based (ChAdOx1 nCov-19 263 and Ad26.COV2.S/Janssen) 251 have been linked to cases of myocarditis [264][265][266] and vaccine-induced prothrombotic immune thrombocytopenia (VITT), 267 respectively. Although rare ( 5 per million), VITT due to antibodies to platelet factor 4 typically occurs following a single vaccine dose and may be fatal in some (pulmonary embolus, cerebral venous thrombosis). In contrast, myocarditis, potentially due to an autoimmune response (triggered by molecular mimicry between spike protein and selfantigen), tends to be more common after the second mRNA vaccine dose but is comparatively less life-threatening as the majority of cases spontaneously resolves. [264][265][266] Future direction Current evidence for the treatment of long COVID is lacking, though many clinical trials for long COVID and CV sequelae ( Table 2) are currently underway. A list of selected clinical trials in long COVID are presented in Supplementary material online, Table S2 to demonstrate the diversity of treatments under investigation. Studies include a variety of rehabilitation programmes (telemedicine and face-face) for treatment of fatigue, cognitive decline and breathlessness, therapies targeted at cognition (e.g. transcranial stimulation), metabolic modulators (e.g. niagen), immunomodulatory therapies (e.g. steroids, laranilubmab, tocilizumab, atorvastatin, colchicine), antifibrotic treatments (e.g. pirfenidone, LYT-100), and anticoagulation (e.g. apixaban). The World Health Organization (see https://clinicaltrials.gov/ct2/who_table or https://www.who.int/ clinical-trials-registry-platform) and ClinicalTrials.gov (see https:// clinicaltrials.gov/ct2/results?cond=COVID-19) list more than 730 studies related to COVID-19; .80 have a major emphasis on longterm CV outcomes. As examples, selected studies with 400 or more participants are listed in Table 3.

Research priorities
Current priorities for research include (i) establishing the prevalence of persistent or chronic SARS-CoV-2 induced CV injury; (ii) elucidating causal mechanisms including the role of immune system, obesity, endotheliopathy, and genetic predispositions; (iii) developing an understanding of CV symptom burden (as part of the long COVID spectrum) and its association with pathology; (iv) developing and refining scalable diagnostic methods with high specificity for COVID-19-associated CV complications (including POTS); (v) identifying novel therapeutic solutions or repurposing old drugs that can protect or reverse COVID-19-associated long-term CV injury; (vi) evaluating the role of vaccination and SARS-CoV-2 variants on cardiac injury; (vii) evaluating the long-term impact of SARS-CoV-2 infection on those with pre-existing cardiac diseases and future risk of heart failure, ischaemic events, and arrhythmias; (viii) evaluating the effects of COVID-19-related autonomic dysfunction on CV homeostasis; and (ix) understanding the impact of long COVID on healthcare costs and on working population. 4,230,233 A concerning trend observed in recent studies 36,19,121,268 is the marked dissociation seen between symptoms and objective measures of health highlighting the limitations of routine clinical investigations. In this regard, deeper phenotyping efforts including advanced cardiopulmonary imaging (e.g. hyperpolarized xenon 269 ), mass spectrometry, 270 metabolomics, 271 proteomics, 272 whole-genome sequencing, and gut microbiome 273 studies promise to unscramble the Rubix cube of pathophysiological processes that underpin long COVID. It is anticipated that multiple endophenotypes 230,233 (inflammatory, metabolic, autoimmune, neurocognitive, psychological 233 ) will surface, through artificial intelligence-driven data analysis, enabling precision diagnostics, prognostic risk prediction, and therapeutics for patients.

Conclusion
Long COVID is emerging as a major public health issue. Our current understanding of pathophysiological mechanisms and treatment options remains limited; however, there is great optimism  Long COVID/post-acute sequelae of COVID-19 as several national and international research initiatives promise to disentangle the complexities of this disease. The high burden of cardiopulmonary symptoms along with other organ manifestations underscores the need for multispecialty input, 274,275 a model that is likely to also profit other chronic diseases. Proactive screening and investigation, where appropriate, could allay fears and anxiety among patients. Considerable efforts to find the right balance between cost-effective investigations and benefit to patients are needed to ensure sustainable service provision in these challenging economic times. Finally, the vast inequalities 43,276 in healthcare provision exposed by COVID-19 will continue to be magnified by long COVID, a problem that calls for global humanitarian efforts to promote and fund equitable access to healthcare, social and welfare support, and vaccines across the world.

Supplementary material
Supplementary material is available at European Heart Journal online. is also a nonexecutive director and consultant to Caristo Diagnostics and holds two patents (one granted and one in review) on pericoronary fat assessment for risk prediction. T.F.L. has no conflicts of interest related to this manuscript but received educational and research grants from Abbott, Amgen, AstraZeneca, Boehringer Ingelheim, Daichi-Sankyo, Novartis, Sanofi, Servier, and Vifor outside this field. D.A.B. has no conflicts of interest related to this manuscript but has received consultant fees from Bayer.