Appraising cardiac dysfunction in liver transplantation: an ongoing challenge


  • Ahmed Zaky,

    Corresponding author
    1. Department of Anesthesiology and Critical Care Medicine, University of Alabama at Birmingham, Birmingham, AL, USA
    • Correspondence

      Ahmed Zaky, MD, MPH

      Assistant Professor of Anesthesiology and Critical Care Medicine, University of Alabama at Birmingham,

      826 Jefferson Tower,

      625 19th St. South, Birmingham, AL, USA

      Tel: 205 934 5272

      Fax: 205 934 2564


    Search for more papers by this author
  • Karim Bendjelid

    1. Department of Anaesthesiology, Pharmacology and Intensive Care, Geneva, Switzerland
    Search for more papers by this author


End-stage liver disease (ESLD) is a multisystemic disease that adversely and mutually aggravates other organs such as the heart. Cardiac dysfunction in ESLD encompasses a spectrum of disease that could be aggravated, precipitated or be occurring hand-in-hand with coexisting aetiological factors precipitating cirrhosis. Additionally and more complexly, liver transplantation, the curative modality of ESLD, is responsible for additional intra- and postoperative short- and long-term cardiac morbidity. The phenotypic distinction of the different forms of cardiac dysfunction in ESLD albeit important prognostically and therapeutically is not allowed by the current societal recommendations, due to conceptual, and methodological limitations in the appraisal of cardiac function and structure in ESLD and in designing studies that are based on this appraisal. This review comprehensively discusses the spectrum of cardiac dysfunction in ESLD, discusses the limitations of the current appraisal of cardiac dysfunction in ESLD, and proposes a hypothetical approach for studying cardiac dysfunction in liver transplant candidates.


coronary artery bypass grafting


cirrhosis-associated cardiomyopathy


coronary artery disease


ejection fraction


end-stage liver disease


international normalized ratio


mean pulmonary artery pressure


orthotopic liver transplantation


portopulmonary hypertension


tranesophageal echocardiography


transpulmonary gradient


transthoracic echocardiography

End-stage liver disease (ESLD) is a multisystem disease that complexly and mutually interacts with other body organs. The heart is one of the organs most adversely affected by liver disease both directly and indirectly. Cardiac dysfunction in the setting of cirrhosis may contribute to mortality as high as 50% post-liver transplantation [1]. The spectrum of heart diseases associated with cirrhosis includes three major groups:

  • Heart disease aggravated by cirrhosis
  • Disease that is caused by a pathological process that concomitantly affects the heart and the liver
  • Cardiac disease (vascular, myocardial or pericardial)

Liver transplantation while considered the definitive treatment of patients with ESLD, can independently contribute to further deterioration of pre-existing cirrhosis-associated cardiac dysfunction. These adverse effects occur as a result of acute changes in loading conditions, and the liberation of inflammatory cytokines and other mediators during graft reperfusion [2]. Furthermore, following liver transplantation, there is an increased risk of adverse cardiac events associated with chronic immunosuppressive therapy. Thus, patients such as these warrant a thorough cardiac evaluation prior to being deemed acceptable liver transplant candidates.

A thorough cardiac evaluation of liver transplant candidates is a challenging task, however. Altered cardiac response to stress, heterogeneity and progression of cardiac disease in liver transplant candidates, and the paucity of well-designed studies investigating pre-operative cardiac testing; all explain the current lack of agreement on a single best screening strategy to optimize perioperative outcomes.

This review will discuss the following: profiles of cardiac dysfunction in ESLD, short- and long-term cardiac dysfunction associated with liver transplantation, and the pre-operative evaluation of liver transplant candidates in the light of the current evidence, appraising its limitations. Also, this review will propose avenues for future investigation of cardiac function in liver transplant candidates.

Profiles of cardiac dysfunction in end-stage liver disease

Cirrhosis-aggravated cardiac disease

Coronary artery disease (CAD)

The prevalence of CAD in patients with ESLD is equal to or greater than its prevalence in the general population, particularly among diabetics, with the range being between 2.5% and 27%, depending on the selection criteria. Reports on the impact of CAD on outcomes following liver transplantation have varied chronologically. A previous report by Plotkin et al. [1] reported an overall mortality of 50% over a 1–3-year follow-up period involving thirty-two liver transplant recipients with angiographically proven CAD. This high mortality rate occurred irrespective of the treatment modality for CAD. Interestingly, a recent multicentre cohort study by Wray et al. [3] reported no significant differences in survival between liver transplant recipients with angiographically proven obstructive CAD (≥50% stenosis) and those with non-obstructive CAD. Of intrigue, the obstructive coronary artery group demonstrated no statistically significant survival difference between patients with severe (>70% stenosis) and those with moderate CAD (50–70% stenosis). This apparent disagreement between the two studies can be explained by several factors. Firstly, both studies reported on patients with coronary artery interventions prior to liver transplantation, but were not specifically designed to assess the effects of these interventions. Moreover, it could be that the significant improvement of management of coronary risk factors over the time between both studies may have contributed to the observed differences. Secondly, the study by Plotkin et al. included only 32 liver transplant candidates and did not include a control group. Thirdly, the study by Wray et al. suffered from several limitations including: the retrospective nature of the analysis leading to lack of adjustment for other covariates that affect survival, lack of adjustment for the interactions between centres and patient groups, lack of data on patients who underwent coronary interventions and did not undergo liver transplantation, and the risk of a classification bias based on coclassification of patients with non-obstructive CAD and normal coronaries as a single group. Taken together, more controlled studies are needed to assess the effects of different coronary interventions on the outcomes after liver transplantation.

Risk stratification of patients with CAD undergoing liver transplantation

Multiple traditional and non-traditional cardiac risk factors have been identified in patients suffering from cirrhosis. Liver transplant candidates with more than one cardiac risk factor such as age >60, male gender, history of CAD, dyslipidaemia, smoking and diabetes have been associated with significant CAD [4]. Of these risk factors, diabetes mellitus (DM) has been specifically associated with asymptomatic coronary artery disease, with concomitant Type I DM, and CAD associated with 40% lower 5- year patient and graft survival rates [5]. Several non-traditional cardiac risk factors have been identified in patients with ESLD. In a recent retrospective case-–control study, patients with non-alcoholic steatohepatitis (NASH) were more likely to be older, be morbidly obese, have hypertension, or have dyslipidaemia compared with alcoholic cirrhosis. On multivariate analysis, NASH was associated with the development of a cardiovascular event (death from any cardiac cause, myocardial infarction, arrhythmias, stroke, cardiac arrest and/or acute heart failure), independent of traditional risk cardiac risk, after liver transplantation [6], compared with patients with alcoholic cirrhosis. There were no significant differences between both groups in overall mortality, however. The association between NASH and traditional risk factors for CAD has been confirmed in other studies [7, 8]. Prospective studies are needed to establish the role of NASH on cardiac-related events after liver transplantation. Multiple other non-traditional risk factors for CAD such as concomitant renal failure [9], elevated C-reactive protein [10] and intracoronary calcium [11] have been studied in patients with cirrhosis. The role of these factors as predictors vs. as markers of cardiac-related morbidity and mortality remains to be investigated in larger studies.

Pre-operative evaluation of patients with risk factors for CAD (Fig. 1)

The evaluation of cirrhosis patients with risk factors for CAD is a major clinical dilemma. This is because non-invasive cardiac testing has a low predictive value for angiographically detected obstructive CAD [12]. In addition, CAD that is caused by non-obstructive atherosclerotic plaques (<50% stenosis) which are not detected by non-invasive testing, may be responsible for acute coronary syndromes (unstable angina, myocardial infarction and sudden cardiac death) [13]. To further complicate issues, patients with myocardial perfusion defects and no angiographically detected coronary lesions may succumb to post-liver transplantation because of cardiovascular complications, primary graft failure and sepsis [14]. It is therefore crucial to exercise caution in interpreting a stress test in liver transplant candidates, even in the absence of angiographical evidence of coronary obstruction. To follow is a brief description of the diagnostic tests employed in the evaluation of liver transplant candidates with risk factors for CAD.

Figure 1.

Schematic diagram of pre-operative evaluation of patients with CAD undergoing OLT.CAD, coronary artery disease; OLT, orthotopicliver transplantation; CABG, coronary artery bypass grafting.

Stress testing

The most recent report from the American Heart Association [15] recommends performing a non-invasive cardiac stress test on patients with ESLD with >3 risk factors for CAD (age >60, diabetes, hypertension, left ventricular hypertrophy, history of CAD, smoking and dyslipidaemia), regardless of functional status. Patients with proven ischaemia on stress testing are then referred for coronary angiography to definitively evaluate their coronary anatomy.

Stress testing describes the physical or pharmacologically induced increase in myocardial oxygen demand to uncover areas of critically obstructed coronary artery disease. Ischaemic myocardial areas manifest either as wall motion abnormalities detected by echocardiography or perfusion defects detected by radionuclide myocardial perfusion scans. Pharmacological stress may be induced by dobutamine, an inotrope, and a chronotrope or by coronary vasodilators such as adenosine or dipyridamole. The latter produces a ‘coronary steal’ from areas of coronary obstruction to areas of normal reactive coronary anatomy and hence precipitates ischaemia. The more significant the coronary artery obstruction, the more predictive the stress test. Stress testing, however, cannot identify non-obstructing atherosclerotic plaques even if they are morphologically vulnerable to rupture. Of stress tests currently being utilized, dobutamine stress echocardiography (DSE) has been the most studied.

Dobutamine stress echocardiography has produced inconclusive results in a large percentage of liver transplant candidates because of the inability to achieve predicted target heart rates or heart rate blood pressure product that are high enough to induce wall motion abnormalities. This physiological limitation is attributed to failure of beta receptors to respond to sympathetic stimulation, a condition termed chronotropic incompetence [16]. Reports on the accuracy of DSE varied widely among studies as a result of the variability in patient selection criteria to perform DSE and end points used for its evaluation. In a retrospective study [12] of 105 patients with end-stage liver disease who underwent both DSE and coronary angiography, Harinstein et al. reported a low sensitivity (13%) and intermediate negative predictive value (75%) for the detection of angiography-proven obstructive CAD. Interestingly, the specificity and positive predictive values of DSE increased significantly with the increase in the number of risk factors for CAD. A recent single-centred retrospective study [17] of 400 liver transplant candidates reported a fairly high NPV (89%) and a low sensitivity (14%) of pre-operative DSE in predicting 30-day post-liver transplantation cardiac events (death/non-fatal myocardial infarctions). Of interest, none of the traditional risk factors for CAD were shown to predict the study outcomes on multivariate analysis. Furthermore, patients in these studies had low Model for End-Stage Liver Disease (MELD) scores. The strong NPV of pre-operative DSE for myocardial injury post-liver transplantation was confirmed in another retrospective study [18]. Taken together, a conclusively negative DSE (lack of wall motion abnormalities on reaching 85% of predicted heart rate response after dobutamine injection) in patients with low cardiac risk may reliably rule out ischaemia or the development of cardiac events post-liver transplantation. A positive test on the other hand, may not necessarily be associated with these outcomes.

Assessment of myocardial perfusion using intravenous ultrasound contrast agents during real-time myocardial stress echocardiography (RTMSE) has been introduced for the assessment and prediction of cardiac events in patients with advanced liver disease. This technique, by combining wall motion abnormalities with perfusion defects is theorized to be a more sensitive measure of myocardial ischaemia than DSE. In addition, adding a contrast agent to echocardiography improves image resolution and endocardial border detection, and thus overcomes two of the major limitations of DSE [19]. Tsutsui et al. [20] examined for the first time the prognostic power of myocardial perfusion imaging with RTMSE in patients undergoing OLT. An abnormal myocardial perfusion imaging was shown to be an independent predictor of in-hospital and 2-year mortality. Interestingly, adjusting for myocardial perfusion, wall motion abnormalities lost predicative value, denoting the higher predictive power of perfusion compared to wall motion abnormalities. Despite its appeal, myocardial perfusion with RTMSE needs to be further studied before being used routinely in liver transplant candidates.

Reports on myocardial perfusion stress testing in the form of nuclear single-photon emission computed tomography (SPECT) with the injection of vasodilators such as adenosine or dipyridamole are conflicting and inconclusive [21-23]. One explanation for these conflicting results is the failure of adenosine or dipyridamole in achieving coronary vasodilation because of the underlying systemic vasodilation in advanced liver disease.

Coronary calcification

Calcium commonly deposits on coronary atherosclerotic plaques. Thus, the degree of plaque calcification can be used to assess the severity of coronary obstruction, and hence the propensity of developing cardiac events after liver transplantation. Computerized tomography is used to measure coronary calcification and the amount of cardiac calcium detected is fit into scores that take into account the age and gender. Calcium scores can thus grade the severity of CAD. Cardiac calcium scores have a limited predictive value as a single screening test for CAD [24]. Cardiac calcium scoring may provide additive information to other CAD risk factors for the occurrence of cardiac events (death, myocardial infarction or revascularization) [25]. It is also of value as a strong negative test; where cardiac calcium scores of zero are associated with < 50% coronary artery stenosis [26]. A recent study of 101 patients with ESLD showed a strong relationship between cardiac calcium score and multiple specific cardiac risk factors (age, systolic blood pressure, fasting glucose levels, some elements of metabolic syndrome). In this study, the cardiac calcium score revealed a high prevalence of asymptomatic CAD in patients undergoing assessment for liver transplantation [11]. Taken together, the role of calcium cardiac score is auxiliary at best. More studies are needed to explore its role in the prediction of cardiac events in patients with end-stage liver disease.

Coronary angiography

Coronary angiography can be performed despite coagulopathy, although at the price of increased risks of bleeding and contrast-induced nephropathy [27, 28]. Sharma and colleagues [28] investigated the safety of cardiac catheterization in patients awaiting liver transplantation and compared them with matched controls with normal liver function. Patients in the study group had a significantly elevated serum creatinine, lower haemoglobin levels and higher international normalized ratios (INR) compared with matched controls. Compared with controls, the study group had a significantly higher rate of vascular complications, transfusions and major bleedings [28]. A transradial approach to coronary angiography has been suggested as a potential approach to decrease bleeding complications in ESLD patients [29]. Also, cardiac CT angiography is emerging as a non-invasive alternative to coronary angiography for the assessment of CAD in patients with ESLD [30]. However, more studies are needed to establish its role in pre-transplant evaluation.

Aspects of treatment of liver transplant candidates with cardiac risk factors

There is no current consensus on what grade of coronary lesion requires intervention and what intervention might be the most useful in patients with ESLD. Coronary revascularization should be considered in liver transplant candidates when the lack of revascularization presents a prohibitive risk factor for liver transplantation. Available options are: medical management, percutaneous revascularization and coronary artery bypass grafting (CABG).

Medical management

Aspirin in the doses used for cardiovascular protection may affect platelet function and worsen thrombocytopaenia commonly occurring in patients with ESLD [31]. The use of angiotensin-converting enzyme inhibitors and angiotensin receptor blockers may worsen renal function in patients with ESLD especially those suffering from hepatorenal syndrome, as well as worsening the systemic vasodilation characteristic of ESLD [32].

The use of statins has raised concerns of worsening liver function [33]. Recent data, however suggest their safety in patients with advanced liver disease provided there is adequate monitoring of liver function [34]. Some concerns have arisen regarding reports of severe rhabdomyolysis with the concomitant use of statins and some antirejection medications that inhibit hepatic cytochrome p-450 enzyme [35]. Adequately powered randomized trials are needed regarding the safety of the use of statins in ESLD patients.

Beta-blockers have been shown to be protective against the occurrence of cardiac events post-liver transplantation. A recent observational study has shown a protective role of beta-blockers in cardiac patients with ESLD [17]. Non-selective beta-blockers are of proven benefit in ESLD patients with oesophageal varices [36]. However, more trials are needed to determine the dosage and timing of initiation of beta-blockers as cardio-protective agents in liver transplant candidates.

Percutaneous revascularization

There is a tendency to favour the use of bare metal stents (BMS) over drug-eluting stents (DES) to limit the duration of dual antiplatelet therapy in the form of aspirin and clopidogrel in the setting of coagulopathy frequently encountered in patients with ESLD. This, however, may curtail the beneficial effects of DES especially in situations when the risk of restenosis from BMS is deemed to be too high; patients with diabetes, complex lesions (bifurcations and left main), long lesions in small calibre vessels and the treatment of in-stent restenosis [37].

Coronary artery bypass grafting (CABG)

Coronary artery bypass grafting surgery is associated with significant morbidity and mortality in patients with ESLD compared with those without ESLD. Early reports suggest a mortality of 30–40% and morbidity of up to 70–100% [38]. Studies have consistently shown a proportionately high mortality and morbidity (including but not limited to: ascites, cardiac tamponade, pleural effusion, bleeding and encephalopathy) with worsening Child and MELD classification in patients undergoing CABG prior to liver transplantation [39-42]. A ‘cut-off’ Child score >7 was found to have a sensitivity and specificity of 86% and 92%, respectively for mortality with a PPV and NPV of 67% and 97%, respectively in some studies. Some studies have reported less mortality with off pump CABG (OPCAB) in patients with advanced Child classification [43]. Several pre- and intra-operative predictors of morbidity and mortality were identified in ESLD patients undergoing CABG; low platelet counts, higher MELD scores, lower serum cholinesterase levels, cardiopulmonary bypass time, and aortic cross-clamp time [44]. Taken together, lower cirrhosis severity scores are associated with lower morbidity and mortality compared with more advanced scores. In the latter group of patients, OPCAB may be offered, however at the expense of an elevated risk of mortality and morbidity.

Combined CABG and liver transplantation has been scarcely reported in the literature. In a case series of patients with severe triple vessel disease and ESLD, there were no intra-operative deaths. At 25 months of follow-up, there was an 80% graft- and patient-survival (one patient died of recurrent Hepatitis C). Overall average length of stay was 21 days [45].

Cardiac surgery can be performed after liver transplantation with an acceptable risk [46]. In a case series of 15 liver transplant recipients who underwent cardiac surgery, there were no early deaths, no rejection episodes or increase in wound infection rates. At a mean of 26.5 months of follow-up, 13.3% had died, and 255 developed recurrent angina [46].

Systemic diseases affecting the heart and the liver

Alcoholic heart muscle disease

Reported incidences of alcoholic cardiomyopathy range from 21 to 32% of dilated cardiomyopathy in surveys conducted at referral centres [47]. Heavy alcohol drinking (>90 g/d) may lead to cardiac structural abnormalities such as: increased cardiomyocyte apoptotic cell death, decreased calcium sensitivity, decreased myofibril function, and myocardial fibrosis [48-50].

Clinical features

The clinical symptoms and signs are similar to heart failure because of other causes. Echocardiographical findings include: decreased ejection fraction, impaired mean fractional shortening, increased left ventricular mass, increase left atrial dimensions, all manifesting as impaired both systolic and diastolic -function. In advanced cases, echocardiography shows: 4-chamber dilatation, normal or decreased left ventricular wall thickness. Supraventricular arrhythmias are common in alcoholic cardiomyopathy. Causes of death are progressive chronic heart failure and sudden cardiac death [50].


Abstinence from alcohol in the early stages may lead to significant improvement of left ventricular ejection fraction [51]. The treatment of alcoholic cardiomyopathy is similar to heart failure of other aetiologies. End-stage heart failure is an indication for cardiac transplantation. Unfortunately, alcohol relapses are common [52].


Amyloidosis describes the multisystem or localized interstitial deposition of different types of proteinaceous materials. Cardiac involvement varies with the type of amyloid material deposited and portends a grave prognosis. Cardiac involvement may manifest as restrictive cardiomyopathy with elevated filling pressures, small-vessel coronary artery disease with persistent elevation of cardiac troponins, arrhythmias in the form of atrial fibrillation, atrio-ventricular blockade, and ventricular fibrillation that may be fatal [53]. Electrocardiography, echocardiography and cardiac magnetic resonance imaging are valuable diagnostic tools in diagnosing and quantifying cardiac amyloidoses.

Treatment of cardiac amyloidoses consists of supportive, modulatory and radical treatments. Supportive treatment includes maintaining higher filling pressures, judicious use of angiotensin-converting enzyme inhibitors, diuretics and beta-blockers, and treatment of arrhythmias that may require insertion of heart rhythm management devices such as biventricular pacemakers and implantable cardioverters-defibrillators. Digoxin is less tolerated as it interacts with amyloid material and deposits in the myocardium. Peripheral oedema may respond to the use of compression stockings. Modulatory treatment consists of various types of chemotherapeutic agents guided by the type of amyloid protein deposits. Radical treatment is used with severe end organ involvement and includes orthotopic or stem cell transplantation [54].

Iron overload cardiomyopathy

Iron overload cardiomyopathy describes the primary or secondary accumulation of iron resulting from either an increased iron absorption or increased parenteral iron intake because of either hereditary or acquired diseases of iron metabolizing proteins.

The heart, liver and endocrine systems are the major organs offended by the disease. Cardiac involvement may be in the form of dilated or restrictive cardiomyopathy leading to systolic or diastolic dysfunction respectively. Cardiac arrhythmias, atrial and/or ventricular, may follow. Involvement of the cardiac conduction system is less common. Cardiac magnetic resonance imaging is of special utility in quantifying cardiac muscle involvement and in monitoring therapy. Echocardiography is used to diagnose the different phenotypic expressions of cardiac muscle overload [55].

Treatment of iron overload cardiomyopathy involves supportive therapy of heart failure syndromes and arrhythmias and the use of chelation therapy to prevent further iron deposition. Organ transplantation is the last resort for end-stage heart and liver disease.

Cirrhosis-associated cardiac diseases

Cirrhosis-associated cardiomyopathy (CAC)

Cirrhosis-associated cardiomyopathy describes cardiac dysfunction in patients with cirrhosis characterized by impaired contractile responsiveness to stress and/or altered diastolic function with electrophysiological abnormalities in the absence of other known cardiac disease [56]. This syndrome is seen in approximately 40–50% of adult patients with cirrhosis [56]. The clinical manifestations of this syndrome correlate with the severity of cirrhosis [57].

Clinical characterization of cirrhosis-associated cardiomyopathy

Cardiac dysfunction resulting from cirrhosis encompass: systolic dysfunction, diastolic dysfunction and electrophysiological dysfunction (Table 1).

Table 1. Working definition of cirrhosis-associated cardiomyopathy
Cardiac dysfunction in patients suffering from cirrhosis characterized by impaired contractile responsiveness to stress and/or altered diastolic relaxation with associated electrophysiological abnormalities in the absence of other known cardiac disease
Diagnostic criteria
Systolic dysfunction
Blunted increase in cardiac output with exercise, volume challenge or pharmacological stimuli
Resting EF < 55%
Diastolic dysfunction
E/A < 1
Prolonged deceleration time (>200 msec)
Prolonged isovolumetric relaxation time (<80 msec)
Supportive criteria
Electrophysiological abnormalities
Chronotropic incompetence
Electromechanical uncoupling
Prolonged QTc interval
Enlarged left atrium
Increased myocardial mass
Increased BNP, pro-BNP
Increased Troponin I
Systolic dysfunction

According to the current working definition of cirrhosis-associated cardiomyopathy, systolic dysfunction describes a contractile defect that is uncovered by stress. The high resting cardiac output and lower filling pressures encountered in patients with cirrhosis is partially explained by low systemic vascular resistance and increased arterial compliance [58]. Physical exercise, however, is associated with a significant elevation of left ventricular filling pressures and a relatively smaller increase in cardiac output, ejection fraction and heart rate [59]. A less than optimal exercise-induced increase in ejection fraction in the presence of an exercise-induced lowering of afterload is a sign of left ventricular contractile dysfunction. Recently, cardiac magnetic resonance imaging (CMRI) and advanced echocardiography technologies have uncovered more of the subtleties of CAC. As assessed by CMRI, there is a modest increase in left ventricular mass, left ventricular end-diastolic and left atrial volumes [60]. A recent echocardiographical study using tissue Doppler imaging [61] revealed a significant increase in left ventricular end-diastolic diameter and a reduction in peak systolic velocity and systolic strain rate. Peak left ventricular systolic velocity and strain rate measured by tissue Doppler are considered more sensitive indices of left ventricular contractile function than the ejection fraction and cardiac index [62]. This is because these markers are more indicative of the more vulnerable longitudinally arranged subendocardial fibres [62]. These findings strongly suggest resting structural and contractile changes in patients with cirrhosis that are not included in the current definition.

Reduced systolic function of both ventricles may have prognostic implications such as the development of ascites and renal dysfunction [56].

Taken together, evidence exists demonstrating that contractile dysfunction in cirrhosis takes place under resting conditions and that it has prognostic implications.

Diastolic dysfunction

Diastolic dysfunction is a prominent feature of CAC. This describes impairment of ventricular filling as a result of alterations in the receptive ventricular properties.

The underlying mechanism of diastolic dysfunction in cirrhosis is increased myocardial wall stiffness most likely because of myocardial hypertrophy, fibrosis and subendothelial oedema [58] resulting in high filling pressures of the left ventricle and atrium and ultimately increasing the risk of pulmonary oedema because of backward failure.

Normally, the velocity of early rapid ventricular filling (denoted by E) is greater than the late filling phase that is dependent on atrial contraction (denoted by A) [63]. Therefore, E/A less than 1 may denote impaired ventricular relaxation. A low E/A, however, is highly preload-sensitive. The American Society of Echocardiography has included tissue Doppler imaging criteria in the diagnosis of diastolic dysfunction [64]. Doppler tissue imaging measures the slow velocity high amplitude annular tissue motion (denoted by E’), which is less, affected by preload. An increase in the E/E ratio has been used as a more sensitive measure of diastolic dysfunction [65].

Electrophysiological dysfunction

Electrophysiological dysfunction includes: prolonged QT interval, chronotropic incompetence and electromechanical dissociation. QT interval prolongation adjusted for heart rate (QTc) is found in approximately 50% of patients with cirrhosis [66]. It has been shown to be significantly related to the severity of liver disease, plasma norepinephrine levels and the presence of portal hypertension [67]. Prolonged QT interval was independently associated with the risk of sudden death in cirrhosis, although the latter is relatively uncommon [66]. The most likely underlying mechanism is dysfunctional potassium channels prolonging the duration of action potential and QT interval [68]. Studies on the dispersion of QT interval (i.e. difference between the longest and shortest interval) report a normally maintained diurnal variation in patients with cirrhosis [69]. A recent retrospective study reported that prolonged QTc (>463 msec) independently predicted mortality from gastrointestinal bleeding [70]. Beta receptor blockade seems to restore a normal QT interval in some individuals [71]. Also, prolonged QTc is partly reversible by liver transplantation, despite being prolonged during the early phase of transplantation [72].

Chronotropic incompetence refers to the inability of the heart rate to respond to physiological and pharmacological demands such as exercise, head tilt, inotropes, and a higher than normal increase in plasma norepinephrine concentrations. The inability to increase the heart rate in response to demands may partially explain reduced cardiac output under these conditions. The time between the onset of electrical and mechanical systole is normally tightly controlled and is referred to as electromechanical coupling. A defect in electromechanical coupling leads to the dys-synchrony between electrical and mechanical systole. Bernardi et al. [73] demonstrated prolongation of pre-ejection phase at rest, together with defective shortening with exercise in patients with cirrhosis denoting a defect in electromechanical coupling.

Aspects of treatment

To date, there are no clinical trials on the management of CAC. Patients with heart failure should be treated following guidelines for non-cirrhosis-induced cardiac failure. Noteworthy, the use of afterload reducers may not be well-tolerated given the widespread and progressive vasodilation characteristic of cirrhosis. Short courses of non-selective beta-blockers were shown to restore prolonged QT intervals towards normal [71]. No recommendation, however, for the chronic use of beta-blockers can be made at the present time. Cardiac glycosides are less effective inotropes. A potential role for the aldosterone antagonist, K-canrenoate exists to reverse the RAAS-induced myocardial fibrosis in pre-ascitic cirrhosis [74]. Despite the theoretic appeal, more studies are needed to further explore this approach. Liver transplantation is associated with a recovery of CAC.

Portopulmonary hypertension (POPH)

Portopulmonary hypertension is a form of pulmonary arterial hypertension (PAH) with increased pulmonary vascular resistance (PVR) or transpulmonary gradient (TPG) because of vasoconstriction and pulmonary vascular remodelling, all occurring in the setting of a clinical evidence of portal hypertension [75] (Tables 2, 3).

Table 2. Definition of portopulmonary hypertension
  1. MPAP, mean pulmonary artery pressure; PVR, pulmonary vascular resistance; PCWP, pulmonary capillary wedge pressure.

Clinical evidence of portal hypertension, and:
MPAP > 25 mmHg at rest or >30 mmHg on exercise
PVR > 240−5
PCWP < 15 mmHg or TPG > 12 mmHg
Table 3. Classification of portopulmonary hypertension
SeverityMean PAP (mm Hg)
  1. PAP, pulmonary artery pressure.

Mild>25 and <35
Moderate>35 and <45
Severe≥ 45

It should be emphasized that portal hypertension is in itself associated with a hyperdynamic circulation, a high transpulmonary flow, sodium retention and volume overload, all can cause PAH with normal TPG and PVR, and hence not classified as POPH. Portopulmonary hypertension is present in about 5%–10% of liver transplant candidates [76]. Approximately 5% of liver transplant candidates have moderate to severe PAH (mean pulmonary artery pressure mPAP >35 mmHg), which has been considered a contraindication to OLT [77]. A pre-operative mPAP 35–50 mmHg has been associated with a 50% risk of mortality after liver transplantation [78]. An mPAP of >50 has been associated with nearly 100% mortality [78]. It is to be emphasized that right ventricular function is more important prognostically than absolute pulmonary artery values, as right ventricular function is the most important criterion in determining the likelihood of a successful graft [79]. Patients with severe POPH may undergo combined heart–lung transplantation.

Diagnostic evaluation
Transthoracic echocardiography (TTE)

Transthoracic echocardiography with Doppler estimation of pulmonary artery systolic pressure (PASP) is the screening tool of choice for the detection of POPH [80]. An elevated PASP >30 mmHg is suggestive of pulmonary hypertension. Additional finding on echocardiography include an enlarged or dilated right ventricle and evidence of right ventricular strain.

The accuracy of an echocardiography derived PASP of >30 mmHg as a threshold cut-off for the diagnosis of POPH was associated with a sensitivity and specificity of 100% and 96%, respectively, among 165 liver transplant candidates. PPV and NPV at this cut-off were 59% and 100% respectively [81].

Pulmonary artery systolic pressure is determined based on an estimate of right atrial pressure and peak tricuspid regurgitation velocity using the modified Bernoulli equation. Right atrial pressure can be estimated from inferior vena caval diameter as elegantly described by Bendjelid et al. [82].

Given the low PPV and the poor correlation between TTE and catheterization values, all patients with echocardiographical findings suggestive of pulmonary hypertension should be referred to right cardiac catheterization for definitive diagnosis.

Right heart catheterization

Right heart catheterization is considered the gold standard for the diagnosis of POPH. It is important to rule out elevated PAP because of elevated left atrial pressures or to elevated cardiac output associated with ESLD. Measurement of the TPG is often used to achieve such a differentiation [83]. Also, a trial of diuresis can be given if PAOP is >15 mmHg with subsequent measurement of mPAP. If mPAP is >35 mmHg, PVR >3 Woods Units and PAOP <15 mmHg, then mild POPH is diagnosed [84].

Radiological studies

Chest radiography and computed tomography may show signs of right ventricular enlargement. Chest CT may also show increased pulmonary artery diameter, and increased segmental artery to bronchial ratios in multiple pulmonary lobes [83].


Signs of right heart strain; right axis deviation, right bundle branch block and first-degree atrio-ventricular block may be seen [85].


Pulmonary vasodilators have been used in moderate and severe POPH (mPAP >35 mmHg) to lower mPAP and hence facilitate liver transplantation [86]. There is a paucity of randomized trials on the use of specific pulmonary vasodilators in patients with POPH. The current treatment centres on the use of prostanoids, endothelin antagonists and phosphodiesterase V inhibitors. Prostanoids induce pulmonary vasodilation and inhibit platelet aggregation leading to an increased risk of bleeding and systemic hypotension respectively. Endothelin antagonists inhibit the endothelin-1-induced vasoconstriction and vascular remodelling. Bosenten, an oral dual endothelin receptor antagonist, carries an FDA caution against its use in patients with moderate to severe liver dysfunction or elevated transaminases. Multiple reports have shown improvement in haemodynamics, exercise capacity with Bosentan with up to three-fold increase in serum transaminases greater than the upper level of normal that responded to dose reduction or discontinuation [87, 88]. Phosphodiesterase V inhibitors enhance nitric oxide-mediated vasodilation. They have been used successfully in small studies in variable doses [89]. Given the paucity of evidence, pulmonary vasodilators should be used with caution in liver transplant candidates.

Liver transplantation may [90] or may not [91] reverse POHP. Interestingly, there have been recent reports of de novo pulmonary hypertension post-OLT. Unfortunately, this syndrome is associated with high mortality with no pre-operative clinical predictors for its development [92].

Hepato-pulmonary syndrome (HPS)

Hepato-pulmonary syndrome is defined as the triad of liver disease, hypoxia that is worse on acquiring an erect posture, and evidence of intrapulmonary vasodilation [93]. Hepato-pulmonary syndrome has been diagnosed in non-cirrhotic liver diseased patients as well as in those without pulmonary hypertension.

The patho-physiology of HPS centres on the release of endogenous pulmonary vasodilators such nitric oxide (NO) and carbon monoxide (CO) via different mechanisms. It is postulated that liver injury leads to the release of endothelin-1 (ET-1), which in turn up-regulates ET-b receptors on pulmonary vascular endothelium leading to the release of endothelial NO (eNO). The latter diffuses into the pulmonary vascular smooth muscle cells to mediate pulmonary vasodilation and arterial-venous shunting. In addition, the release of tumour necrosis factor alpha (TNF-α) in response to endotoxaemia resulting from gut-translocated bacteria leads to up-regulation and release of inducible NO (i-NO), which contributes to pulmonary vasodilation. Bacterial translocation also triggers phagocytosis by alveolar macrophages that result in alveolar capillary angiogenesis and intrapulmonary shunting [94].

Diagnostic evaluation
Arterial blood gases

An (A-a) gradient of >15 mm Hg (>20 mmHg if age > 64 years) while a patient is breathing room air in the sitting position constitutes the diagnosis of hypoxia in patients with liver disease.

Demonstration of intrapulmonary shunting: echocardiography

The appearance of bubbles of agitated saline in the pulmonary veins for five cardiac cycles after the injection of agitated saline in an upper extremity is one test to prove the presence of intrapulmonary shunting. Normally, these bubbles are trapped within pulmonary capillaries. Their appearance in pulmonary veins signifies the presence of pulmonary vasodilation and intrapulmonary shunting. Importantly, the appearance of these bubbles earlier than five cardiac cycles raises suspicion of intracardiac shunting.

A less sensitive yet more specific test of intrapulmonary vasodilation is the demonstration of radiolabelled albumin particles (MAA) that are normally trapped in the lungs, in the brain and kidney vasculature following injection in a peripheral vein, and total body scanning.

Of the two tests, echocardiographic bubble studies are used for routine screening for PPS in liver transplant candidates.


Liver transplantation constitutes the only effective treatment of PPS. Conflicting data [95, 96] exist on the correlation of the degree of pre-operative hypoxaemia and mortality post-liver transplantation.

Pericardial disease

Fluid retention characteristic of ESLD frequently leads to ascites, pleural and pericardial effusions, of which the latter may be associated with cardiac tamponade. The diagnosis of cardiac tamponade may be challenging in patients with cirrhosis. This is because elevated right-sided pressures may mask or delay right ventricular (and/or atrial) collapse despite the presence of underlying tamponade physiology [97]. Significant effusions should be treated by pericardiocentesis or pericardial window, and patients should be followed up by echocardiography.

Consequences of cardiac dysfunction on other organ systems in end-stage liver disease

Cirrhosis-associated cardiac dysfunction can gravely and viciously aggravate other organ systems in patients with ESLD. This vicious cycle of adverse consequences adds to the baseline pathological circulatory state encountered in ESLD.

The circulatory state in ESLD is described as; hyperkinetic, centrally under-filled, peripherally vasodilated and mal-distributive. The release of nitric oxide together with failure of the liver to clear circulating vasodilators such as glucagon result in severe splanchnic vasodilation leading to renal and gut under-perfusion.

As a result, sympathetic nervous system, renin–angiotensin system and argenine–vasopressin systems get stimulated. This results in tachycardia and increase in cardiac output, in renal and brain vasoconstriction, and in sodium and water retention, as an attempt to restore mean perfusion pressure to these organs. Because of a surplus of splanchnic vasodilators, the splanchnic circulation becomes more resistant to circulating vasoconstrictors (angiotensin, catecholamines) resulting in vicious release of this vasoconstrictors leading to renal and brain vasoconstriction and the development of ascites [98] (relative hypovolaemia).

As time progresses, the increase in cardiac output becomes insufficient to meet the increased demands in ESLD leading to a relative reduction in effective circulatory volume and perfusion pressures leading to kidney injury. It is thus conceivable that cirrhosis-associated (or -aggravated) cardiac dysfunction can hasten the development of kidney injury and ascites formation. Patients with ESLD who are on beta-blockers are particularly at risk of the negative consequences of cardiac output reduction. In fact beta- blockers (propranolol) were shown to be independently associated with reduced survival in a cohort of cirrhotics with ascites [99]. Similar considerations are applicable for angiotensin-converting enzyme inhibitors which can worsen the underlying severe vasodilatory state in patients with ESLD leading to decreased renal perfusion pressure.

Both systolic and diastolic dysfunctions are associated with kidney injury in ESLD. Right or left ventricular systolic dysfunction may lead to a reduction in the stroke volume and renal blood flow. Backward failure resulting from uni- or biventricular systolic or diastolic dysfunction may result in renal congestion that may add to the renal insult.

Based on the above, it is not surprising what recent literature suggests of a prognostic cardio-renal relationship in ESLD. In a small prospective study, Krag et al. [100] demonstrated that the reduction in cardiac index to below 1.5 L/m/m2 – as measured by myocardial perfusion scan – was associated with the development of type-1 hepatorenal syndrome and that their combination was associated with a significant reduction in survival. Interestingly, cardiac index performed superiorly to MELD score in predicting survival. An earlier study on patients with ESLD and a diagnosis of spontaneous bacterial peritonitis demonstrated a significant inverse correlation between cardiac output and the development of renal failure. It remains to be determined, though, whether cardiac dysfunction is a risk factor vs. a risk marker for hepatorenal syndrome. Also, it remains to be determined whether more sensitive markers of ventricular mechanical dysfunction and renal dysfunction retain a similar relationship compared with their less sensitive counterparts.

Cardiac stress precipitated by liver transplantation

Intra-operative cardiac complications

Orthotopic liver transplantation (OLT) imposes major cardiovascular stresses on the heart that adversely affect pre-existing cardiac dysfunction. Hypothetically, OLT is divided into three stages; pre-anhepatic, anhepatic and neohepatic stages.

The pre-anhepatic stage which starts from surgical incision until cross-clamping of the supra- and infra-hepatic inferior vena cava (IVC) and hepatic artery, is characterized by a decrease in preload and cardiac output. Reduction in preload is attributed to drainage of ascites following surgical incision, and bleeding resulting from pre-existing coagulopathy and dissection. IVC cross-clamping may result in up to a 50% reduction in venous return to the heart. Veno-venous bypass and piggy-back clamping of the IVC are known strategies to ameliorate the impact of preload reduction resulting from IVC cross-clamping [101].

The anhepatic phase is heralded by explantation of the diseased liver and ends with graft reperfusion. During this phase, major coagulopathy (primarily fibrinolysis), acid–base (metabolic acidaemia), and electrolyte disturbances (hypocalcaemia, hypophosphataemia and hypomagnesaemia), hypothermia and hypotension secondary to preload reduction are major morbidities. Coagulation factor replacement, blood transfusion via dedicated rapid infusion systems, central rewarming are all strategies to ameliorate these morbidities [101].

The neohepatic phase represents the most challenging period of cardiovascular instability during OLT. It extends from graft reperfusion until the completion of arterial and biliary anastomosis. Post-reperfusion syndrome, the hallmark of the neohepatic phase, is defined as the decrease in mean systemic blood pressure by greater than 30% for more than 1 min for the first 5 min after graft reperfusion. Typically, post-reperfusion syndrome is associated with myocardial depression, pulmonary arterial hypertension, and significant reduction in systemic vascular resistance and bradycardia. The most plausible theory for its development is the release of cold acidotic hyperkalaemic fluid into the circulation on graft reperfusion. Reactive oxygen species and multiple inflammatory mediators such as tumour necrosis factor alpha are responsible for the clinical phenotype of PRS [102, 103]. These mediators are incriminated in myocardial depression, as well as a severe reduction in afterload [104]. PRS occurs in up to 30% of OLT recipients and may be of enough severity to result in cardiac arrest. Fatal arrhythmias secondary to hyperkalaemia may occur during this phase and warrant prophylactic treatment with hyperventilation, beta-agonists, sodium bicarbonate and insulin prior to graft reperfusion. Catecholamines, vasopressin and methylene blue have all been used in the treatment of hypotension because of PRS. Cardiopulmonary bypass has been instituted successfully in cardiac arrest cases resulting from PRS [105].

The rapidity of development and the impact of these changes on the heart require the institution of invasive real-time cardiovascular monitoring such as arterial line and pulmonary artery catheters with continuous cardiac output and central venous saturation monitoring capability, and transesophageal echocardiography during the procedure.

Post-operative cardiac complications

Post-operative cardiac complications have emerged as a major cause of non-graft related mortality after OLT. The incidence of post-operative cardiac events ranges between 12% and 70% [17, 106]. A possible explanation for this disparity is the considerable variability in the definition of cardiac events, heterogeneity of cardiac risk factors in the studied population and the timing and duration of follow-up of cardiac events post-OLT. For simplicity, cardiac events post-OLT will be discussed in terms of short- and long-term events with 3 months being the cut-off.

Short-term complications after liver transplantation

The immediate post-transplant period is associated with restoration of normal systemic and splanchnic circulations. As a result, systemic vascular resistance increases abruptly and considerably leading to systemic hypertension. The initiation of immunosuppressive agents such as steroids and calcineurin inhibitors may also be associated with hypertension (see below). Abrupt increase in systemic blood pressure may lead to precipitation of heart failure and pulmonary oedema as well demand ischaemia and infarction.

Pulmonary oedema occurs in approximately 47% of patients post-OLT [107]. Fluid overload, increase in afterload and inflammatory capillary leak are all potential causes of post-OLT pulmonary oedema. Compared with hydrostatic type pulmonary oedema, permeability type pulmonary oedema lasts longer and is associated with more prolonged mechanical ventilation, intensive care unit and hospital stays [108]. Judicious fluid management in the post-transplant period is crucial to prevent the occurrence of pulmonary oedema and subsequent adverse outcomes [109]. In haemodynamically stable patients, a negative fluid balance guided by a low central venous pressure may reduce graft congestion, right ventricular dysfunction and arrhythmias.

New-onset heart failure has been recognized post-OLT. In a recent large retrospective study [110], predominant systolic heart failure occurred within the first 60 days post-operatively. Non-ischaemic aetiology was the predominant aetiology. Pre-transplant hypertension, diabetes, mean arterial pressure <60 mmHg, mean pulmonary artery pressure >30 mmHg, mean pulmonary capillary pressure >15, haemodialysis and prolonged QT interval (>450 ms) and elevated brain natriuretic peptide were independent predictors for the development of new-onset heart post-OLT. Interestingly beta-blockers and tacrolimus were protective against the development of systolic heat failure. New-onset reversible dilated cardiomyopathy has been reported in patients with no prior cardiac disease post-OLT [111].

Venous thromboembolism, though rare, has been reported post-OLT. In a single centre report [112], the incidence of venous thromboembolism after OLT was 4.6% (comparable to other major abdominal procedure). History of diabetes, deep vein thrombosis, end-stage renal disease were all predictors of the development of this condition. Unless contraindicated, patients post-OLT should receive appropriate antithrombotic prophylaxis against deep vein thrombosis.

Long-term cardiac complications after liver transplantation

Immunosuppressive agents while reducing the risk of graft rejection carry the potential to induce other morbidities such as diabetes, hypertension, hyperlipidaemia and obesity; all are known risk factors for atherosclerotic coronary artery disease. Clinicians should weigh the risk of rejection against the risk of developing these morbidities by using the least toxic dose of immunosuppressive agents together with lifestyle modifications and the use of medications with least pharmacological interactions. To follow is a brief review of some of the long-term morbidities induced by immunosuppressive agents.

Hypertrophic cardiomyopathy

Tacrolimus has been incriminated in the development of a new-onset hypertrophic cardiomyopathy in a dose-related fashion [113]. Tacrolimus levels >15 ng/dl were most commonly reported to be associated with hypertrophic cardiomyopathy on 2-dimensional echocardiography. Interestingly, cardiomyopathy was reversible with discontinuation of tacrolimus or switching to a different immunosuppressant. Cyclosporine has been incriminated in hypertrophic cardiomyopathy though to a much lesser degree than tacrolimus [114].


Calcineurin inhibitors such as cyclospsorin and tacrolimus are associated new-onset hypertension post-OLT [115]. Mechanisms include; activation of renin–angiotensin pathway, inducible nitric oxide inhibition and activation of endothelin leading to systemic hypertension. Treatment included life style modification, angiotensin-converting enzyme inhibitors with or without calcium channel- and beta-blockers [115]. Lastly, in non-responsive cases, reduced dosages or switching to another family of immunosuppressants may be warranted.

Diabetes mellitus

Corticosteroids, tacrolimus and to a lesser extent cyclosporins are associated with the development of new-onset diabetes. Postulated mechanisms include increased hepatic gluconeogenesis and insulin resistance. Treatment of diabetes goes in line with non-immunosuppression induced diabetes such as life style modifications, use of oral hypoglycaemics and using the least toxic immunosuppressant dosages.


Sirolimus [116] and glucocorticoids are associated with higher prevalence of hyperlipidaemia and hypertriglyceridaemia compared with cyclosporin and tacrolimus. Postulated mechanisms include stimulation of adipose tissue lipases and inhibition of lipoprotein lipases resulting in an increase in hepatic synthesis of triglycerides and very low density lipoproteins. Corticosteroids also activate acetyl coenzyme A carboxylase, free fatty acid synthase and 3-hydroxy-3-methylglutaryl-coenzyme reductase leading to increased triglycerides, and very low density lipoproteins and decreased levels of high density lipoproteins [115]. In addition to life style modification, pharmacological treatment with statins and gemfibrozil is often prescribed. Statins should be prescribed at the smallest effective doses given the inhibitory effects of cyclosporine and tacrolimus on the p-450 enzyme system involved statin metabolism [115].

Taken together, OLT while reversing some of the cardiac morbidities associated with ESLD, it may aggravate and add others. This calls for the need to develop prediction models for their occurrence.

Limitations of the current appraisal of cardiac dysfunction in end-stage liver disease

A successful risk stratification of a systemic disease relies on two steps; firstly on adequate characterization of the disease, and secondly, on careful assessment of its operative risk.

The current cardiac pre-operative evaluation of ESLD candidates suffers from limitations in both steps. As a result, current models of pre-operative risk stratification fail to predict and optimize perioperative outcomes.

To follow is a brief discussion of the some of the limitations in characterization and operative risk assessment.

Limitations in characterization of cardiac dysfunction in end-stage liver disease

The current characterization of cardiac dysfunction in ESLD conceptualizes the heart as a uniform global mechanical haemodynamic pump operated by the left ventricle. As such, global mechanical indices such as cardiac index and ejection fraction are relied upon in the characterization of cardiac function in ESLD. Some of the inherent limitations of these indices are listed.

Cardiac index is a load-determined index of ventricular function. It is not an intrinsic measure of ventricular contraction or relaxation. In addition, cardiac index does not provide information about the function of other chambers of the heart, such as the atria and pericardium. It does not provide information about myocardial tissue motion or size; the latter being relevant functional and prognostic indices of ventricular function. Using this flow index as a marker of ventricular contractility [117, 118] conforms to the conceptual modelling of the heart as a ‘hydraulic pump’ (Table 1).

Ejection fraction (EF), another global index of ventricular ejection, reflects the circumferential shortening of the left ventricle; in this regard, it does not take into account the longitudinally arranged subendocardial fibres that are more prone to ischaemia and fibrosis earlier than the circumferentially arranged fibres [119]. Furthermore, EF is a load-dependent index that does not reflect the intrinsic myocardial contractile function [120, 121], which shares this limitation with cardiac index. It is therefore possible that EF varies across studies, depending on the amount of resuscitation patients have received (preload) and their stage of sepsis (afterload), rather than on a true contractile deficiency. EF being a global index of ventricular function could be insensitive to regional myocardial abnormalities. Hyperkinetic segments bordering hypokinetic segments may thus erroneously produce a normal EF. In this respect, the use of EF as a surrogate of ventricular contractility [122-126] conforms conceptually to regarding the heart as a ‘hemodynamic pump’, not taking into account the intrinsic mechanical function of the heart. Moreover, transesophageal echocardiography (TEE) underestimates ventricular volumes in 10% of the cases [127]. In addition, EF measurement using echocardiography depends on an accurate identification of the endocardial borders, which could be inaccurate in approximately 30% of the cases when the images assessed are of low resolution [128]. Transthoracic echocardiography (TTE) may be less optimally performed in critically ill patients, who receive mechanical ventilation, have chest tubes or have increased body habitus [121]. Notably, TEE may underestimate ventricular volumes compared to TTE and transpulmonary thermodilution [127].

In contradistinction to the current conceptualization, the heart is an embryologically heterogeneous organ that is inherently and intrinsically load-interactive [129]. Models of cardiac structure [130, 131] have agreed on a multilayered heterogeneous arrangement of myocardial fibres. Consequently, the cardiac motion is reciprocal during both systole and diastole. Meanwhile, this reciprocal heterogeneity is interactive with both pre- and afterload. This fact argues against the assumption of clear-cut phases of the cardiac cycle, where the heart only and uniformly contracts or relaxes. It also suggests that load-interaction should be considered an intrinsic function of the heart. Accordingly, ‘load-independent’ indices of cardiac performance may not be a physiologically appropriate term to describe cardiac performance.

Limitations in the assessment of perioperative cardiac risk

The current literature lacks a definitive pre-operative strategy that optimizes perioperative outcomes in liver transplant candidates. This stems from the lack of randomized controlled trials investigating the utility of pre-operative testing in high-risk liver transplant candidates. In addition, there is a lack of consensus on what risk factors are associated with post-operative cardiac dysfunction. Adding to the difficulty, the literature is inconclusive on what constitutes a cardiac morbidity post-OLT, let alone the contribution of pre-operative risk factors. Moreover, it yet remains to be proven whether new-onset cardiac morbidities (secondary to immunosuppression, as an example) are truly ‘new-onset’ vs. ‘newly-diagnosed’. It is therefore understandable why the current guidelines for pre-operative evaluation of liver transplant candidates do not stand on solid evidence-based recommendations. To follow are some examples of the current literature on cardiac testing in liver transplant candidates.

In a retrospective study, Ripoll et al. ([132].) observed that patients who developed an abnormal cardiac response defined by a coupled reduction in left ventricular stroke work index with an elevation of filling pressures to be more likely hyponatraemic, and centrally hypovolaemic. An abnormal cardiac response occurred most commonly after reperfusion and was associated with prolonged intubation. Although adding to the notion that an abnormal cardiac response may occur in the absence of an abnormal pre-operative echocardiographic cardiac testing, this study is criticized for being single centred, retrospective, lacking a control group, using non-conventional limits for statistical significance, and using a preload-determined global marker of left ventricular systolic dysfunction. In the absence of a baseline testing of what the author defined as an abnormal cardiac response, it is hard to agree to the conclusion that the attained response was exclusively because of ESLD. Another retrospective report [133] commented on intra-operative and 6 months post-operative cardiac morbidity. Cardiac morbidity was defined as pulmonary oedema, arrhythmia, ischaemia, pulmonary hypertension and heart failure. Using multivariate regression analysis, the authors concluded that a high integrated-MELD score, occurrence of intra-operative cardiac complications, pre-operative history of cardiac disease were all independent predictors of poor cardiac outcomes. Interestingly, serum sodium shown to be associated with worse outcomes in Ripoll's study was not shown to be associated with poor outcome in this study. This study is criticized based on its retrospective nature, lack of a control group, missing relevant patient data, not reporting on the diagnostic modality used to determine baseline cardiac dysfunction, of only adjusting for age and gender in the multivariate regression model, and in not reporting baseline characteristics comorbidities of the study cohort. Based on these limitations, firm conclusions on specific pre-operative cardiac predictors are hard to be generalized.

Given the above-mentioned inherent limitations of dobutamine stress echocardiography in the pre-operative evaluation of liver transplant candidates, it is not surprising that its use has produced mixed results. Umphry et al ([134].) reported that a failure to attain a heart rate greater than 82% of predicted and a maximum rate pressure product of 63.3 to be predictive of poor intra- and post-operative cardiac outcomes. On the other hand, Findlay et al. [18] reported more disappointing results by demonstrating failure of a positive pre-operative dobutamine stress echocardiogram to predict cardiac morbidity in the form of myocardial injury. Again both studies suffer from the above-mentioned limitations of being retrospective and of reporting different definitions of cardiac outcomes at different time points in a small sample-sized single-centred population. Therefore, there is a need for a standardized pre-operative approach for risk stratifying liver transplant candidates and that this approach be prospectively tested.

Future directions

There has been a recent interest in identifying cardiac dysfunction at a subclinical level. Advanced echocardiographic techniques and magnetic resonance imaging are now gaining more popularity in detecting early mechanical dysfunction of the myocardium that is not detected by conventional echocardiographic techniques such as the ejection fraction.

Contrast magnetic resonance imaging has been shown to be superior to echocardiography in detecting smaller size perfusion defects not manifesting as wall motion abnormalities by echocardiography [135]. This is theorized to be because of a better detection of the extent of perfusion in the longitudinally arranged subendocardial fibres that are more prone to ischaemia compared with circularly and radially arranged fibres.

In an interesting exploratory study, Reddy et al. [136] demonstrated feasibility of what they called one-stop shop for cardiac, abdominal and vascular MRI in liver transplant candidates. More studies are needed to investigate the influence on outcomes of various techniques of cardiac MRI in the pre-operative cardiac evaluation of liver transplant candidates.

One potential modality to detect subtle myocardial dysfunction in ESLD is to study the regional deformation of the left ventricle along different planes of motion. In this regard, the present assessment could be achieved by measuring strain and strain rate using advanced echocardiographic techniques, such as Doppler tissue imaging and speckle tracking echocardiography. Strain measures segmental myocardial deformation and strain rate measures the rate at which this deformation occurs [137]. Measuring strain across the longitudinal plane of motion has been shown to be a more sensitive measure of ischaemia, fibrosis and hypertrophy compared to more global indicators of left ventricular systolic function, such as ejection fraction [137]. This is, in part, because of the arrangement of the more vulnerable subendocardial fibres of the left ventricle across this plane of motion. Doppler tissue imaging used to measure myocardial velocities is angle-dependent and cannot be performed in retrospect. Speckle tracking imaging creates reflections and interferences between the ultrasound beam and the myocardial tissue ‘speckles’ that can be tracked retrospectively throughout the cardiac cycle and is angle-independent [138]. In a recent study in a paediatric population with septic shock, speckle tracking imaging modality was able to detect impaired myocardial performance that was not revealed by EF measurement [139]. Speckle tracking echocardiography may thus be a modality to explore in the pre-operative evaluation of liver transplant candidates.

Considerations to approach pretransplant work-up (Fig. 2)

Characterization of cardiac dysfunction in liver transplant candidates should be go hand- in-hand with risk stratification. Cardiac characterization should take into consideration the heterogeneous nature of the heart and should focus on segmental rather than global indices. The proposed approach should assess the heart electrically, mechanically and metabolically aiming at identifying early markers of cardiac dysfunction. Meanwhile risk stratification should start by identifying the interaction between cirrhosis and baseline cardiac disease. This interaction is important in identifying whether cardiac dysfunction is cirrhosis-aggravated vs. cirrhosis-induced. Based on this classification, and in the presence of identified markers of cardiac dysfunction, a hypothetical scoring system can be designed that aims at stratifying patients undergoing OLT into mild, moderate and high-risk groups. Longitudinal studies are needed with temporally and clearly defined cardiac outcomes. The interaction of chemotherapy with identified pre-operative risk factors is crucial for the prediction of long-term post-operative cardiac outcomes.

Figure 2.

Considerations to approach pretransplant work-up.


The heart and the liver are mutually interactive organs. The current preliver transplant work-up does not stem from solid evidence. As a remedy, standardized definitions of risk profiles and cardiac morbidities should be prospectively investigated both short and long term. This standardization should go in line with the quest of newer modalities capable of diagnosing cardiac dysfunction at an early stage. The involvement of a multidisciplinary team of caregivers is a key to the success of this approach.


Conflict of interest: The authors do not have any disclosures to report.