Clinical predictors of post–liver transplant new-onset heart failure


Address reprint requests to Waqas Qureshi, M.D., Department of Internal Medicine, Henry Ford Hospital/Wayne State University School of Medicine, 2799 West Grand Boulevard, CFP 1, Detroit, MI 48202. Telephone: 313-916-2958; FAX: 313-916-1394;


Objectives of this study were (1) to evaluate preoperative predictors of systolic and diastolic heart failure in patients undergoing liver transplantation (LT) and (2) to describe the prognostic implications of systolic and diastolic heart failure in these patients. The onset of heart failure after orthotopic LT remains poorly understood. Data were obtained for all LT recipients between January 2000 and December 2010. The primary outcome was post-LT heart failure: systolic (ejection fraction ≤ 50%), diastolic, or mixed heart failure. Patients underwent echocardiographic evaluation before and after LT. Pretransplant variables were evaluated as predictors of heart failure with Cox proportional hazards model. 970 LT recipients were followed for 5.3 ± 3.4 years. Ninety-eight patients (10.1%) developed heart failure in the posttransplant period. There were 67 systolic (6.9%), 24 diastolic (2.5%), and 7 mixed systolic/diastolic (0.7%) heart failures. Etiology was ischemic in 18 (18.4%), tachycardia-induced in 8 (8.2%), valvular in 7 (7.1%), alcohol-related in 4 (4.1%), hypertensive heart disease in 3 (3.1%), and nonischemic in majority of patients (59.2%). Pretransplant grade 3 diastolic dysfunction, diabetes, hypertension, mean arterial pressure ≤ 65 mm Hg, mean pulmonary artery pressure ≥ 30 mm Hg, mean pulmonary capillary wedge pressure ≥ 15 mm Hg, hemodialysis, brain natriuretic peptide level and QT interval > 450 ms were found to be predictive for the development of new-onset systolic heart failure. However beta-blocker use before LT and tacrolimus after LT were associated with reduced development of new-onset systolic heart failure. In conclusion, pretransplant risk factors, hemodynamic variables, and echocardiographic variables are important predictors of post-LT heart failure. In patients undergoing LT, postoperative onset of systolic or diastolic heart failure was found to be an independent predictor of mortality. Liver Transpl 19:701–710, 2013. © 2013 AASLD.


late atrial ventricular filling velocity


brain natriuretic peptide


confidence interval


early ventricular filling velocity


transmitral Doppler early filling velocity/tissue Doppler early diastolic mitral annular velocity


hazard ratio


liver transplantation


Model for End-Stage Liver Disease


Model for End-Stage Liver Disease–sodium

Cardiac dysfunction is a well-known but poorly defined entity among patients with cirrhosis.[1] Many of these patients improve after liver transplantation (LT),[2] although there are only limited data in the form of case reports[3, 4] and small, retrospective studies[5, 6] describing new-onset heart failure after LT. Heart failure after LT may account for approximately 7.4% of deaths.[7]

To date, sparse data are available regarding the predictors of heart failure after orthotopic LT. It is also unclear whether patients with preexisting cirrhotic cardiomyopathy are likely to develop heart failure in the posttransplant period. Cirrhotic cardiomyopathy has been shown to deteriorate after periods of stress or certain procedures such as a transjugular intrahepatic portosystemic shunt procedure.[8] The clinical manifestations of heart failure become overt in the immediate posttransplant period.[9, 10] Because of improved overall care, however, LT recipients are living longer, and they develop cardiac complications such as heart failure in the long term.[11] A preoperative evaluation of certain variables may provide both the clinician and the patient with valuable information regarding the risk of posttransplant heart failure and mortality as well as potential implications and, therefore, allow an informed decision regarding LT.

The aims of our study were as follows:

  1. To evaluate the utility of pre-LT hemodynamic, clinical, biochemical, and echocardiographic parameters as predictors of posttransplant heart failure.
  2. To determine the etiology of new-onset heart failure after LT.
  3. To assess new-onset heart failure as a cause of significant mortality in the posttransplant period.


Study Population

Adult patients (>18 years old) who underwent LT at the Henry Ford Health System (Detroit, MI) between January 2000 and December 2010 were selected. The study protocol was approved by the institutional research review board. Patients were identified from the LT registry database of the Henry Ford Health System. A pretransplant cardiac assessment was performed for all patients, and this included an electrocardiogram and a transthoracic echocardiogram. On the basis of the cardiologist's assessment of an individual's risk of coronary artery disease, some patients underwent noninvasive testing such as stress myocardial imaging and perfusion scans, whereas others underwent invasive testing in the form of cardiac catheterization.

Data Collection

The following pretransplant variables were recorded on the basis of a detailed chart review: demographic data (age, sex, race, and body mass index), cardiovascular risk factors (defined later), systolic and diastolic blood pressures, pulse rate, peak troponin level, Model for End-Stage Liver Disease (MELD) score, indication for LT, QTc interval, beta-blocker use, diuretic use, pretransplant echocardiographic variables (described later) and Charlson comorbidity score. The operative time, intraoperative cardiac arrests, and posttransplant immunosuppressant use were also recorded. All patients were admitted to the surgical intensive care unit for standard hemodynamic monitoring after transplantation.


Hypertension was defined as the presence of an elevated systolic blood pressure > 140 mm Hg or a diastolic blood pressure > 90 mm Hg on 2 separate occasions at least 1 week apart or as the use of antihypertensive medications. Diabetes was defined as a fasting blood glucose level > 126 mg/dL or a hemoglobin A1c level > 6.5% on 2 occasions. Hypercholesterolemia was defined as a total cholesterol level > 200 mg/dL, and smoking was categorized as active smoking (within last 6 months), past smoking, or never smoking. QT intervals were corrected for the heart rate with Bazett's equation (QTc = QT/RR1/2). Myocardial infarction was defined as the presence of positive cardiac markers with or without electrocardiogram changes and clinical documentation of myocardial infarction treatment for >48 hours and no evidence of secondary causes of myocardial infarction in the history.


The echocardiographic assessment was performed in accordance with American Society of Echocardiography guidelines. Standard 2-dimensional and Doppler echocardiography was performed. The examination was performed in the left lateral decubitus position by ultrasonography technologists and was interpreted by a cardiologist blinded to the clinical details of the patient. The ejection fraction was assessed with the modified biplane Simpson method in the apical view as well as the 4-chamber view. Doppler echocardiography was used to evaluate the diastolic filling pattern. The mitral inflow velocity curve [the peak early ventricular filling velocity (E)], the peak late atrial ventricular filling velocity (A), the deceleration time, and the E/A ratios were calculated. To measure the early peak velocity, tissue Doppler imaging was used at the level of the lateral mitral annulus. Transmitral Doppler early filling velocity/tissue Doppler early diastolic mitral annular velocity (E/e′) ratios were also calculated. The pulmonary artery systolic pressure was measured. Left and right heart catheterization was performed in a few patients with elevated pulmonary artery pressures on transthoracic echocardiography in order to rule out severe pulmonary hypertension, which is a contraindication to LT. Diastolic dysfunction was defined as shown in Table 1. Pre-LT echocardiographic data were available for all patients, with complete data available for 945 patients (97%). Complete post-LT echocardiographic data were available for only 853 patients (88%).

Table 1. Classification of Left Ventricular Diastolic Dysfunction
Diastolic DysfunctionMitral InflowTissue Doppler Imaging
Normal (grade 0)E/A = 2.3 ± 0.6E/e′ ≤ 10
Abnormal relaxation (grade 1)E/A ≤ 1.0E/e′ ≤ 10
Pseudo-normal (grade 2)E/A = 2.3 ± 0.6E/e′ > 10
Restrictive physiology (grade 3)E/A > 1.5E/e′ > 10


The primary outcome was the development of new-onset heart failure. Heart failure was defined clinically as the presence of at least 2 symptoms of heart failure (paroxysmal nocturnal dyspnea, orthopnea, lower extremity edema, and shortness of breath) in addition to 1 clinical feature (presence of S3, jugular venous distention, and pulmonary rales) and either an elevated brain natriuretic peptide (BNP) level > 100 pg/mL or chest X-ray findings of pulmonary edema. Heart failure was classified via echocardiography as systolic (ejection fraction ≤ 50%), diastolic (as defined in Table 1), or mixed heart failure. The time to the onset of posttransplant heart failure was calculated as the interval between the date of transplantation and the date of the transthoracic echocardiogram. The etiology of heart failure was obtained from a comprehensive review of cardiology documentation, discharge summaries, laboratory tests, biopsy studies, cardiac imaging, echocardiography, and noninvasive and invasive cardiac testing.


The follow-up period began on the date of transplantation and ended on either the last day of follow-up preceding December 31, 2011 or the date of expiration. The date of expiration was obtained from the Social Security database, which was supplemented by our institution's transplant registry in order to ensure no loss to follow-up.

Complete follow-up was ensured for all patients via cross-referencing with the Henry Ford Hospital liver transplant database.

Statistical Analysis

Normally distributed continuous data are expressed as means and standard deviations or, if skewed, as medians and interquartile ranges. Categorical data are expressed as numbers and percentages. Various predictors of posttransplant heart failure were calculated with a Cox proportional hazards model with an adjusted analysis for age, sex, race, etiology, Framingham risk score, and Charlson comorbidity index. The survival analysis was performed with the Kaplan-Meier method for patients who developed cardiac failure. The Gehan-Breslow-Wilcoxon test was used to compare cumulative survival. Predictors of the development of diastolic and systolic heart failure were studied separately. Various patient characteristics, allograft characteristics, and pretransplant cardiac characteristics were studied as predictors of heart failure. A P value  < 0.05 was considered statistically significant.


Incidence of Heart Failure

Nine hundred seventy patients underwent LT at our institution between January 2000 and December 2010 (mean age = 53.2 ± 10 years); 48.5% were women, and 64.5% were white. These patients were followed for a period of 5.3 ± 3.4 years. Among these, 98 patients developed heart failure in the posttransplant period, with systolic heart failure in 67 patients (6.9%), diastolic heart failure in 24 patients (2.5%), and mixed systolic/diastolic heart failure in 7 patients (0.7%); this yielded an overall incidence of 10.1%. The mean pre-LT echo interval was 54 ± 42 days before transplantation, and the post-LT echo interval was 61 ± 41 days after transplantation. The mean Charlson score was 5.8 ± 3.1, and the mean Framingham risk score was 5.2 ± 4.8. BNP was available only for 423 patients (43.6%), and troponin was recorded for 612 patients (63.1%). Both BNP and troponin levels were significantly elevated in patients with heart failure. The baseline characteristics of these patients are provided in Table 2.

Table 2. Baseline Characteristics
VariablesHeart Failure (n = 98)No Heart Failure (n = 872)P Value
  1. a

    Quit more than 1 year ago.

  2. b

    Either quit within 1 year or currently smoking.

Age (years)55.6 ± 7.952.9 ± 10.20.03
Race [n (%)]   
White61 (62.2)565 (64.8)0.66
African American34 (34.7)238 (27.3)0.12
Other3 (3.1)69 (7.9)0.10
Male [n (%)]51 (52.0)449 (51.5)>0.99
Known coronary artery disease [n (%)]7 (7.1)49 (5.6)0.50
Hypertension [n (%)]74 (75.5)511 (58.6)0.001
Diabetes [n (%)]60 (61.2)399 (45.8)0.004
Family history of coronary artery disease [n (%)]18 (18.4)131 (15.0)0.38
Remote smoker [n (%)]a33 (33.7)398 (45.6)0.02
Recent smoker [n (%)]b23 (23.5)127 (14.6)0.03
Depression [n (%)]55 (56.1)598 (68.6)0.02
Dialysis [n (%)]10 (10.2)69 (7.9)0.43
Alcohol/drug use [n (%)]11 (11.2)50 (5.7)0.046
Framingham score12.4 ± 2.87.4 ± 2.5<0.001
Charlson comorbidity score6.2 ± 3.65.5 ± 3.00.05
Psychiatric disorder [n (%)]9 (9.2)58 (6.7%)0.4
Mean MELD score19.6 ± 9.118.2 ± 9.30.14
Medications [n (%)]   
Beta-blocker45 (45.9)288 (33.0)0.01
Diuretic55 (56.1)439 (50.3)0.29
Aldosterone antagonist49 (50.0)413 (47.4)0.67
Immunosuppressive therapy   
Tacrolimus71 (72.4)776 (89.0)<0.001
Mycophenolate mofetil86 (87.8%)804 (92.2)0.17
Steroid94 (95.9)840 (96.3)0.78
Thymoglobulin2 (2.0)15 (1.7)0.69
Echocardiographic parameters   
Left ventricular diastolic dysfunction [n (%)]31 (31.6)78 (8.9)<0.001
Pre-LT ejection fraction (%)56.0 ± 4.557.2 ± 4.80.17
Post-LT ejection fraction37.8 ± 13.760.7 ± 3.9<0.001
Cardiac enzymes   
High-sensitivity troponin I (ng/mL)0.05 ± 0.120.01 ± 0.070.001
BNP (pg/mL)120.3 ± 139.361.5 ± 68.4<0.001
Hemodynamic measurements   
Heart rate (bpm)89 ± 2374 ± 220.21
Mean arterial pressure (mm Hg)84 ± 4291 ± 340.32
Mean portal artery pressure (mm Hg)20.4 ± 4.517.2 ± 4.9<0.001
Right atrial pressure (mm Hg)10.2 ± 5.37.3 ± 6.40.53
Pulmonary artery systolic pressure (mm Hg)34.4 ± 8.626.6 ± 9.1<0.001
Pulmonary capillary wedge pressure (mm Hg)16.6 ± 5.114.2 ± 7.60.18
Cardiac index3.4 ± 2.64.2 ± 1.90.03

The mean donor age was 43.1 ± 17.0 years, and 29% were female. There were 734 donors (75.7%) who were on at least 1 vasopressor medication before organ donation.

The mean warm ischemia time was 43.4 ± 13.1 minutes, and the mean cold ischemia time was 6.6 ± 2.3 hours. The mean hepatic flow that was achieved was 393.5 ± 224 mL/minute, and the mean portal vein flow that was achieved was 778 ± 416 mL/minute. The mean fat content of the donor livers was 5% ± 7.6%.

The aggregate blood products that were transfused intraoperatively included 5.5 ± 8.2 U of packed red blood cells, 9.4 ± 5.3 U of fresh frozen plasma, 1.52 ± 2.1 U of cell saver, and 10.7 ± 8.3 U of cryoprecipitates. Complex arterial reconstruction was performed in 7% of the patients, and a complex biliary procedure was performed in 5% of the patients.

The major indications for LT are given in Table 3. All patients had a diagnosis of cirrhosis before transplantation. This was confirmed by explant pathology. The mean intensive care unit stay was 5 days (range 2-14 days), and the mean length of stay was 12 days (range 8-23 days).

Table 3. Indications for Transplantation
Etiologyn (%)
Hepatitis C444 (45.8)
Alcohol167 (17.2)
Hepatitis B27 (2.8)
Nonalcoholic steatohepatitis44 (4.5)
Cryptogenic76 (7.8)
Hepatocellular carcinoma17 (1.8)
Autoimmune hepatitis27 (2.8)
Drug-induced toxicity9 (0.9)
Primary biliary cirrhosis30 (3.1)
Primary sclerosing cholangitis52 (5.4)
Alpha-1-antitrypsin deficiency2 (0.2)
Acute fatty liver of pregnancy2 (0.2)
Caroli's disease2 (0.2)
Cholestatic liver disease1 (0.1)
Portal venous thrombosis2 (0.2)
Wilson's disease3 (0.3)
Focal nodular hyperplasia1 (0.1)
Mycotic aneurysm1 (0.1)
Secondary biliary cirrhosis1 (0.1)
Polycystic liver disease8 (0.8)
Hemochromatosis10 (1.0)
Other/unclear44 (4.5)

The baseline characteristics indicated that patients with heart failure were more likely to be older, hypertensive, diabetic, alcoholic, and tobacco users (recent or remote smokers). They were also more likely to be on beta-blockers, were less likely to be treated with tacrolimus after transplantation, and had higher levels of high-sensitivity troponin and BNP. These patients were also more likely to have been diagnosed with diastolic dysfunction in the 6 months preceding transplantation.

Etiology and Time to the Onset of Heart Failure

There were 69 patients (70.4%) who were evaluated for an ischemic etiology with cardiac catheterization, whereas 45 patients (45.9%) were evaluated with some form of stress testing after transplantation. Suspected etiologies of heart failure included ischemia in 18 patients (18.4%), tachycardia-induced cardiomyopathy in 8 patients (8.2%), valvular heart disease in 7 patients (7.1%), alcoholic heart disease in 4 patients (4.1%), and hypertensive heart disease in 3 patients (3.1%); the etiology was unclear or nonischemic in the majority of the patients (58 or 59.2%). The median time to the onset of heart failure was 101.5 days (25.5-268 days; Fig. 1). Systolic dysfunction occurred immediately (<7 days) after transplantation in 17 of the 67 cases (25.4%), and 9 of these cases had an intraoperative or acute postoperative myocardial infarction.

Figure 1.

Time to the onset of heart failure.

Predictors of Posttransplant Heart Failure

Medical History

Underlying medical conditions such as hypertension, diabetes, dyslipidemia, a previous smoking history, a history of depression, a family history of coronary artery disease, a history of psychiatric disorders, and an alcohol/cocaine history were not found to be significant risk factors for posttransplant heart failure in a multivariate analysis.


A pretransplant mean arterial blood pressure ≤ 65 mm Hg was associated with posttransplant heart failure [hazard ratio (HR) = 1.83, 95% confidence interval (CI) = 1.07-3.14, P = 0.03]. Pretransplant hemodynamic variables based on right heart cardiac catheterization were available for 379 patients (44 had heart failure). An elevated pulmonary capillary wedge pressure ≥ 15 mm Hg (HR = 5.3, 95% CI = 2.1-13.0, P = 0.0001) and a mean pulmonary artery systolic pressure (HR = 2.71, 95% CI = 1.12-6.53, P = 0.03) were significant predictors of post-LT heart failure. The mean pulmonary artery pressure was 31.9 ± 9.1 mm Hg, and the mean capillary wedge pressure was 14.6 ± 6.1 mm Hg. Low mean arterial pressures (≤65 mm Hg) were present in 119 patients (12.3%) and normalized in 94 patients (79.0%) after transplantation. Posttransplant hemodynamic readings were available for only 101 patients (10.4%). Among these, the pulmonary artery pressure normalized to <30 mm Hg in 74 patients (73.3%), and the pulmonary capillary wedge pressure normalized to <15 mm Hg in 38 patients (37.6%).

Laboratory and Electrocardiographic Parameters

The presence of a higher pretransplant MELD score resulted in a trend toward the development of posttransplant systolic heart failure, but this did not reach statistical significance (HR = 1.02, 95% CI = 1.00-1.05, P = 0.05). Elevated blood urea nitrogen levels, hemodialysis, elevated BNP levels, and prolonged QTc intervals were all predictive for the development of new-onset systolic heart failure in these patients (Table 4). There were 495 patients (51%) with a prolonged QT interval, and this normalized in 312 patients (63.0%) within the first year after transplantation.

Table 4. Pretransplant Variables That Are Predictors of Posttransplant Systolic Heart Failure
PredictorsSystolic Heart Failure [HR (95% CI)]P ValueDiastolic Heart Failure [HR (95% CI)]P Value
  1. a

    There were no patients.

Risk factors    
Hypertension2.08 (1.27-3.38)0.0031.34 (0.99-1.87)0.07
Diabetes1.88 (1.21-2.92)0.0051.82 (1.01-3.27)0.047
Current smoking history0.85 (0.66-1.10)0.211.23 (0.68-2.10)0.75
History of psychiatric disease1.31 (1.05-1.62)0.020.89 (0.65-1.21)0.45
Alcohol use1.98 (1.05-3.75)0.041.47 (0.77-2.93)0.34
Drug use1.15 (0.67-1.97)0.621.21 (0.78-1.98)0.43
Dyslipidemia1.39 (0.79-2.46)0.251.98 (1.11-3.55)0.02
Family history of coronary artery disease1.30 (0.74-2.27)0.361.48 (0.77-2.84)0.24
Body mass index ≥ 30 kg/m21.31 (0.56-2.56)0.191.88 (1.2-2.4)0.002
Heart rate > 90 bpm1.13 (0.60-2.17)0.70.75 (0.33-1.69)0.49
Mean arterial pressure ≤ 65 mm Hg1.83 (1.07-3.14)0.030.69 (0.16-2.94)0.62
Right atrial pressure > 8 mm Hg1.41 (0.64-3.11)0.41.34 (0.48-4.52)0.68
Mean pulmonary artery   pressure ≥ 30 mm Hg2.71 (1.12-6.53)0.031.63 (0.98-2.06)0.09
Pulmonary capillary wedge   pressure ≥ 15 mm Hg5.3 (2.1-13.0)<0.0011.08 (0.94-1.23)0.27
Cardiac index < 2.62.39 (0.55-10.35)0.242.18 (0.41-11.67)0.36
Diastolic heart failure    
Grade 11.52 (0.78-2.94)0.23- 
Grade 21.75 (0.67-4.54)0.25- 
Grade 31.89 (1.11-3.13)0.02- 
Laboratory parameters    
Sodium < 130 mg/dL1.15 (0.53-2.43)0.720.54 (0.17-1.78)0.31
MELD score1.02 (1.00-1.05)0.050.98 (0.95-1.02)0.32
MELD-Na score1.01 (0.99-1.03)0.341.00 (0.99-1.02)0.84
Blood urea nitrogen (mg/dL)3.04 (1.94-4.77)<0.0012.42 (1.14-2.74)0.009
Estimated glomerular filtration   rate < 60 mL/minute1.54 (1.00-2.40)0.051.63 (0.95-3.30)0.07
Dialysis3.7 (1.52-9.00)0.0042.58 (1.01-3.93)0.01
International normalized ratio > 21.27 (0.73-2.20)0.41.22 (0.55-2.69)0.62
Albumin < 3 g/dL1.18 (0.76-1.85)0.460.87 (0.48-1.58)0.64
Total bilirubin (mg/dL)1.00 (0.64-1.57)1.000.95 (0.54-1.68)0.87
Elevated troponins > 0.04 ng/mL3.22 (0.24-50)0.8411.24 (1.21-104.65)0.03
BNP > 50 pg/mL2.04 (1.21-3.70)0.021.01 (0.42-2.43)0.98
Electrocardiogram QTc > 450 ms4.10 (1.89-8.91)0.00011.79 (1.02-3.16)0.04
Beta-blocker0.59 (0.38-0.92)0.020.78 (0.49-1.10)0.15
Diuretic0.94 (0.21-4.21)0.941.27 (0.74-2.18)0.39
Statin0.80 (0.52-1.24)0.320.94 (0.53-1.89)0.82
Tacrolimus0.48 (0.28-0.82)0.0081.34 (1.01-4.71)0.03
Mycophenolate mofetil0.51 (0.25-1.06)0.070.56 (0.22-1.43)0.22
Thymoglobulin1.15 (0.13-10.39)0.9a 
Steroid use0.76 (0.22-2.61)0.660.70 (0.16-3.06)0.64

Medication Use

Beta-blocker use before transplantation was associated with a reduced risk of developing systolic heart failure after transplantation (HR = 0.59, 95% CI = 0.38-0.92, P = 0.02). Tacrolimus use after transplantation was also found to be associated with a lower risk of systolic heart failure but a higher risk of diastolic heart failure (Table 4).

Statins, mycophenolate mofetil, thymoglobulin, steroids, and diuretics were not associated with a risk of developing posttransplant systolic heart failure.

Echocardiographic Evaluation

Diastolic cardiac dysfunction was an independent predictor of posttransplant systolic heart failure. There were 145 patients with diastolic dysfunction. When they were evaluated for various degrees of diastolic dysfunction, grade 3 diastolic dysfunction, a form of severe diastolic impairment, was found to be strongly predictive of posttransplant systolic heart failure (HR = 1.89, 95% CI = 1.11-3.13, P = 0.02). Most of these patients had left ventricular hypertrophy (76%) with a wall thickness ≥ 1.1 cm. When patients with elevated wedge pressures and elevated pulmonary pressures were excluded, there were only 10 patients left, and diastolic dysfunction was still a significant predictor of posttransplant systolic heart failure for these patients.

Mean portal pressures were available for only 31% of the patients. The mean portal pressure was higher in patients with heart failure versus patients who did not develop heart failure (20.4 ± 4.5 versus 17.2 ± 4.9, P = 0.0001). We did not, however, find the mean portal pressure to be a predictor of the development of post-LT systolic heart failure.

Donor, Allograft, and Operative Characteristics

In order to determine whether posttransplant heart failure was a function of the etiology of liver disease (Table 3), donor organ characteristics, or intraoperative conditions, we ran interaction tests. Using a pairwise analysis, we found that alcoholic liver disease, packed red blood cell transfusions ≥ 8 U, fresh frozen plasma transfusions > 10 U, a donor liver fat content > 10%, the operative time, and intraoperative cardiac arrest were significantly associated with posttransplant heart failure. Other characteristics, however, such as donor age, inotrope use, warm ischemia time, cold ischemia time, length of stay, intensive care unit stay, sex of the donor, donor hepatitis C status, urgent LT, and complex surgical reconstructions were not associated with posttransplant heart failure (Table 5).

Table 5. Intraoperative and Allograft Parameters
 Systolic Heart Failure [HR (95% CI)]P ValueDiastolic Heart Failure [HR (95% CI)]P Value
  1. a

    There were no patients with intraoperative cardiac arrest that developed diastolic heart failure subsequently.

Operative parameters    
Intraoperative cardiac arrest16.39 (4.32-62.5)<0.001* 
Intraoperative time1.002 (1.000-1.003)0.0511.005 (1.002-1.007)<0.001
Intraoperative packed red   blood cell transfusion ≥ 8 U2.45 (1.1-3.6)0.024.5 (2.1-7.8)<0.001
Intraoperative fresh frozen   plasma > 10 U1.9 (1.3-2.5)0.011.6 (1.01-2.2)0.04
Allograft parameters    
Fat content > 10%1.5 (0.7-1.9)0.432.4 (1.2-4.7)0.02
Anatomically abnormal allograft1.3 (0.3-3.2)0.691.2 (0.1-29.9)0.87
Hepatitis C–positive1.1 (0.3-4.5) 1.2 (0.5-4.7)0.52
Warm ischemia time > 1 hour1.6 (0.8-2.1)0.291.3 (0.5-2.5)0.51
Cold ischemia time > 8 hours1.7 (0.7-2.5)0.441.4 (0.5-2.9)0.37
Surgical procedure    
Complex arterial construction1.8 (0.4-6.9)0.291.3 (0.3-5.2)0.39
Complex biliary construction1.8 (0.7-3.1)0.231.9 (0.6-4.4)0.26
Urgent procedure2.3 (0.9-3.9)0.101.5 (0.6-2.7)0.45
Use of inotropes1.9 (0.95-2.8)0.161.4 (0.5-1.9)0.67

Is There a Difference Between Predictors of Systolic and Diastolic Heart Failure?

Various predictors of posttransplant systolic heart failure and posttransplant diastolic heart failure were calculated. Hypertension, diabetes, QT prolongation, intraoperative packed red cell transfusions ≥ 8 U, an elevated blood urea nitrogen level ≥ 30 mg/dL, end-stage renal disease on dialysis, and intraoperative fresh frozen plasma transfusions > 10 U were associated with both systolic and diastolic heart failure.

Factors Unique to Systolic Heart Failure

Although a mean pulmonary artery pressure ≥ 30 mm Hg, a mean arterial pressure ≤ 65 mm Hg, an elevated BNP level, an elevated blood urea nitrogen level ≥ 30 mg/dL, end-stage renal disease on dialysis, and alcohol use were associated with an increased risk of systolic heart failure, beta-blocker use and tacrolimus use were associated with a lower risk of systolic heart failure. Predictors of systolic heart failure are presented in Table 4.

Factors Unique to Diastolic Heart Failure

The intraoperative time, tacrolimus use, elevated troponins within the first postoperative month, a body mass index ≥ 30 kg/m2, dyslipidemia, a diagnosis of nonalcoholic steatohepatitis, and an allograft fat content > 10% were associated with an increased risk of the development of diastolic heart failure (but not systolic heart failure), as shown in Tables 4 and 5.


Intraoperative cardiac arrest occurred during 6 LT operations, but only 1 of these patients developed immediate postoperative systolic dysfunction with an ejection fraction of 15%. Four of the remaining 5 patients developed systolic dysfunction over a period of 5 years.

A total of 270 patients (27.8%) died during the mean follow-up period of 5.3 ± 3.4 years. Forty-two patients (42.9%) died in the heart failure group. Posttransplant heart failure was associated with increased all-cause mortality (HR = 2.3, 95% CI = 1.7-2.9, P = 0.0001). Posttransplant systolic heart failure was associated with a significant mortality risk (HR = 1.58, 95% CI = 1.14-2.21, P = 0.007), whereas diastolic heart failure was found to be a weak independent predictor of post-LT mortality (HR = 1.56, 95% CI = 1.01-2.46, P = 0.044). A Kaplan-Meier survival analysis of patients with heart failure is shown in Fig. 2.

Figure 2.

Kaplan-Meier curves demonstrating cumulative survival for (- - -) patients who developed heart failure and (—) patients who did not develop heart failure.


Our study reveals a high incidence of heart failure (particularly systolic heart failure) among LT recipients with a prolonged follow-up duration. We also evaluated multiple predictors of systolic heart failure, some protective and others deleterious. Systolic heart failure was found to be an independent predictor of mortality in this population.

To the best of our knowledge, this is the largest study describing new-onset heart failure in the post-LT setting. We report a total of 98 such cases. This is consistent with previously cited incidence rates, which have varied from 3.4%[12] to 24%[6] and have depended on the size of the study sample, the inclusion/exclusion criteria, and the definition of heart failure used. Previous studies relied on the clinical definition of heart failure, which is subjective and may miss a significant number of cases that may not be clinically apparent. In our study, we elected to use echocardiographic parameters of heart failure in addition to these definitions in order to obtain a more in-depth evaluation of the pathophysiological and hemodynamic processes at play. Almost a quarter of the patients with new-onset heart failure presented within the first 30 days of LT. These observations suggest that the patients may have had a major intraoperative coronary event or had undetected cardiac disease (perhaps unmasked by the stress of surgery and manifesting as immediate posttransplant heart failure). A low pretransplant mean blood pressure was found to be significantly associated with posttransplant heart failure, and this further supports the notion that hemodynamic changes during surgery may have a significant impact on the already weakened heart.

Elevated pretransplant blood urea nitrogen levels and hemodialysis were found to increase the risk of posttransplant heart failure.[13] Although heart failure is a very common complication in patients on dialysis, there is literature to support the idea that the initiation of hemodialysis in patients with chronic kidney disease may reduce venous congestion and the left ventricular mass index and, in turn, reduce the risk of cardiac hospitalizations and improve cardiac outcomes.[14] High blood urea nitrogen levels have previously been associated with an increased risk of heart failure exacerbation.[15]

A prolonged QTc interval is the main electrocardiographic abnormality in patients with cirrhotic cardiomyopathy[16] and is found in up to 50% of patients with cirrhosis.[17] Bernardi et al.[18] showed that a prolonged QTc interval correlates with increased circulating plasma noradrenaline and worse outcomes. However, it remains unclear whether this correlates with worsened outcomes due to heart failure or arrhythmias resulting in sudden cardiac death.[19] In addition, a prolonged QT interval in patients with cirrhosis is independent of the etiology of liver disease[20] and is partly reversible after transplantation.[21]

Beta-blocker use was associated with a reduced risk of systolic heart failure after transplantation. This association was not demonstrated for diastolic heart failure. A higher proportion of patients with heart failure were on beta-blockers before transplantation. This might be due to the unequal number of patients in the heart failure group with evidence of diastolic dysfunction before transplantation. There was a reduced risk for these patients on beta-blockers who developed systolic heart failure eventually. The role of beta-blockade in reducing QT prolongation has been studied in LT recipients.[22, 23] QT prolongation, however, is not an indication for beta-blocker therapy in patients with cirrhosis because life-threatening arrhythmias are uncommon, and there are few data to suggest that their use improves outcomes in patients with advanced cirrhosis.

Pulmonary hypertension was found to be an important determinant of posttransplant heart failure. Currently, severe pulmonary hypertension is a contraindication for LT because perioperative mortality rates approach almost 100% in patients with a mean pulmonary artery pressure > 50 mm Hg.[24] This may be related to an unmasking of cardiac dysfunction after transplantation in patients with elevated pulmonary pressures leading to heart failure and worse outcomes, as shown in our study. However, with advancements in therapy for pulmonary hypertension, pressures can be reduced to safe levels for LT.[25-27] However, the long-term posttransplant cardiac complications in these patients are still unknown because short-term mortality is high.

Diastolic dysfunction is common in patients with advanced cirrhosis and results from myocardial hypertrophy, subendothelial edema, and elevated filling pressures, which lead to stretching of the myocardial fibers and stimulation of the renin-angiotensin system.[28, 29] Pretransplant echocardiographic parameters of diastolic dysfunction, including E/A and E/e′ ratios, were found to be determinants of posttransplant heart failure in our study. This is relevant to the selection of patients for LT. Currently, the pretransplant cardiac assessment is predominantly based on the determination of the ejection fraction, pulmonary artery pressures, and coronary artery disease. Diastolic dysfunction, however (especially grade 3), was found to be associated with a significantly increased risk of posttransplant heart failure and resulted in increased morbidity and mortality. Previous studies included diastolic dysfunction based only on a single assessment of the E/A ratio and did not classify diastolic dysfunction into 3 grades. This is important because milder grades of diastolic dysfunction may reverse after LT,[2, 30] and these patients may safely undergo transplantation, whereas patients with more advanced diastolic dysfunction (grade 3) have a much higher rate of posttransplant heart failure and worse outcomes. With worsening grades of heart failure, patients are likely to have worsening of diastolic dysfunction. Worsened diastolic dysfunction results in elevated backward pressures and, therefore, more delicate hemodynamic profiles that may manifest as clinical heart failure with even the slightest increase in blood pressure. Diastolic dysfunction is also linked to chronic liver disease. In fact, grade 2 diastolic dysfunction is prevalent in patients with advanced liver disease (Child-Pugh class 3).[31]

Patients with cirrhosis are likely to have lower blood pressures. An impaired renin-angiotensin-aldosterone system has been shown to be a major player in the hemodynamics of patients with cirrhosis.[32, 33] These hormones have also been shown to play a role in cardiac remodeling, such as left ventricular hypertrophy.[34] Hypertension is another major player and was prevalent in more than half of our LT population. We found that blood pressure extremes (eg, a lower mean arterial pressure < 65 mm Hg and an elevated blood pressure > 140 mm Hg) were significantly associated with the development of heart failure. We suspect that this was due to already compromised myocardium in these patients due to an impaired hormonal balance. Beta-blockers showed cardioprotective effects in patients with LT. It has been shown that beta-blockers probably have cardioprotective effects by inhibiting aberrant channel gating of the sarcoplasmic Ca2+ release channel, which is inhibited by beta-blockers.[35] Our findings are contrary to the findings of Sersté et al.,[36] who showed that patients with cirrhosis with concomitant ascites and beta-blocker use had poorer survival. This conflict might be due to the change in the hemodynamics after LT, and as shown in other patient populations at risk, beta-blockers have a similar beneficial effect in this population.[37] In our cohort, we found patients on beta-blockers after transplantation to have better survival profiles, and this finding is consistent with previous studies.[38, 39]

Intriguingly, we found tacrolimus to be cardioprotective in these patients when it was restarted after transplantation. It has been shown that tacrolimus limits leukocyte accumulation and may prevent myocardial injury from ischemic/reperfusion injury.[40] In addition, tacrolimus has weaker drug interactions with statins, and this may allow these patients to be treated with low-dose statins.[41, 42] Statins have been shown to prevent the development of heart failure.[43]

On the other hand, tacrolimus was associated with an increased risk of diastolic heart failure. This effect may be explained by the already known tacrolimus effect of inducing cardiac hypertrophy.[44] In a large pediatric LT series, Dehghani et al.[45] showed that it was associated with hypertrophic cardiomyopathy. On the basis of their findings, Dehghani et al. recommended periodic cardiac evaluations for these patients treated with tacrolimus. This study highlights the importance of a preventive strategy for this population. Many of the factors associated with heart failure in our study may be modified either pharmacologically or nonpharmacologically. The appropriate management of chronic medical conditions (eg, diabetes and hypertension) and a patient's commitment to avoid various habits (eg, smoking and alcohol use) may improve the outcomes of these patients and should be a part of post-LT care.

Lastly, posttransplant heart failure was found to be significantly associated with worsened outcomes in our study after we controlled for confounding variables. This may be due to an increased risk of infection, repeated hospitalizations, and effects on the allograft that were beyond the scope of this study and need to be analyzed in future studies.

There are certain limitations to our study. It is a retrospective study based on chart review. Complete echocardiographic data were not available for all patients; however, these data were available for 88% of the patients, and this still constitutes a large number of patients. The ejection fraction evaluation was available for all the patients in the study cohort, and this was the main feature for determining the development of heart failure, with some form of echocardiographic evaluation performed before transplantation. Posttransplant infections were not assessed as causes of morbidity and mortality. We conclude that posttransplant new-onset heart failure is relatively common after LT and significantly increases mortality. Certain pretransplant risk factors may predict the chances of a transplant recipient developing heart failure in the posttransplant period, and this may help in the pretransplant cardiac evaluation.