Patients transplanted for nonalcoholic steatohepatitis are at increased risk for postoperative cardiovascular events

Authors


  • Potential conflict of interest: Nothing to report.

  • This research was supported by investigator-initiated funds.

Abstract

Nonalcoholic steatohepatitis (NASH) is an independent predictor of coronary artery disease (CAD). Our aim was to compare the incidence of cardiovascular (CV) events between patients transplanted for NASH and alcohol (ETOH)-induced cirrhosis. This is a retrospective cohort study (August 1993 to March 2010) of 242 patients (115 NASH and 127 ETOH) with ≥12 months follow-up after liver transplantation (LT). Those with hepatocellular carcinoma or coexisting liver diseases were excluded. Kaplan-Meier's and Cox's proportional hazard analyses were conducted to compare survival. Logistic regression was used to calculate the likelihood of CV events, defined as death from any cardiac cause, myocardial infarction, acute heart failure, cardiac arrest, arrhythmia, complete heart block, and/or stroke requiring hospitalization <1 year after LT. Patients in the NASH group were older (58.4 versus 53.3 years) and were more likely to be female (45% versus 18%; P < 0.001). They were more likely to be morbidly obese (32% versus 9%), have dyslipidemia (25% versus 6%), or have hypertension (53% versus 38%; P < 0.01). On multivariate analysis, NASH patients were more likely to have a CV event <1 year after LT, compared to ETOH patients, even after controlling for recipient age, sex, smoking status, pretransplant diabetes, CV disease, and the presence of metabolic syndrome (26% versus 8%; odds ratio = 4.12; 95% confidence interval = 1.91-8.90). The majority (70%) of events occurred in the perioperative period, and the occurrence of a CV event was associated with a 50% overall mortality. However, there were no differences in patient, graft, or CV mortality between groups. Conclusions: CV complications are common after LT, and NASH patients are at increased risk independent of traditional cardiac risk factors, though this did not affect overall mortality. (HEPATOLOGY 2012;56:1741–1750)

Cardiovascular (CV) complications are the leading cause of non-graft-related mortality in patients after liver transplantation (LT).1 LT itself imposes stresses on the CV system, such as increased blood pressure and peripheral vascular resistance after transplantation, which can unmask latent or clinically mild cirrhotic myocardial dysfunction, known as “cirrhotic cardiomyopathy”.2, 3 CV disease is also closely associated with metabolic syndrome. Nonalcoholic fatty liver disease (NAFLD) represents the hepatic manifestation of metabolic syndrome. NAFLD represents a wide spectrum, ranging from isolated hepatic steatosis to nonalcoholic steatohepatitis (NASH), that can eventually lead to cirrhosis requiring LT.4

NAFLD is a risk factor for CV disease, independent of metabolic syndrome.5-7 Hepatic steatosis is associated with a proatherogenic lipid profile and increased production of proinflammatory cytokines, including interleukin-6.8-10 In addition, experimental and epidemiological data have shown that CV disease is strongly associated with elevated serum aminotransfereses, a surrogate marker of hepatic necroinflammation.11, 12 Existing data demonstrate that 5-year survival for NASH patients undergoing LT is comparable to other causes of liver disease. However, early mortality is higher in NASH patients, primarily as a result of infectious and CV complications.13, 14 Alcohol (ETOH) is a common etiology of cirrhosis with comparable post-transplant survival outcomes to NASH.13 The underlying cardiac status of patients with cirrhosis is important in determining who will cope with the CV stress of LT. Advanced liver disease leads to metabolic disturbances that alter lipid, glucose, and blood-pressure levels, and certain liver diseases, such as hepatitis C virus (HCV), are associated with an elevated diabetes risk.15 Given the increased metabolic comorbidities found in NASH LT candidates, it would be reasonable to assume they may be at increased risk for postoperative CV complications. Yet, it is unknown whether the presence of NASH is an independent risk factor, aside from traditional CV risk factors, for post-LT CV events. Furthermore, current methods of CV risk assessment may be insufficient to risk stratify patients in this population.

Therefore, the aim of this study was to compare CV morbidity and mortality in patients transplanted for NASH cirrhosis versus those transplanted for ETOH cirrhosis. We hypothesized that NASH patients would have an increased incidence of CV events within 1 year of transplant and that this may lead to decreased survival. Findings from this study will further our understanding of postoperative CV risk in the increasing numbers of NASH patients undergoing LT.

Abbreviations

ATP, Adult Treatment Panel; BMI, body mass index; CAD, coronary artery disease; CI, confidence interval; CLD, chronic liver disease; CV, cardiovascular event; DM, diabetes mellitus; DSE, dobutamine stress echocardiography; ECG, electrocardiogram; EDT, E wave deceleration time; ETOH, alcohol-induced cirrhosis; HCV, hepatitis C virus; HDL, high-density lipoprotein; HL, hyperlipidemia; HR, hazard ratio; HTN, hypertension; LT, liver transplantation; MELD, Model for End-Stage Liver Disease; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; QTc, corrected QT interval; OR, odds ratio; TGs, triglycerides.

Patients and Methods

Patients.

We performed a retrospective review of data in all patients who underwent LT at Northwestern Memorial Hospital (Chicago, IL) and the University of Chicago (Chicago, Il) medical centers from August 1, 1993 through March 1, 2010. LT was performed in 1,273 patients (973 Northwestern, 300 University of Chicago). Of these, 234 (18.3%) had a pretransplant diagnosis of NASH or cryptogenic cirrhosis with known risk factors for NASH. Seventy-nine patients were excluded because of missing data (31 patients) or other contributing causes of liver disease (48 patients). Of the 155 evaluable subjects, 40 patients with active malignancy were excluded. The remaining 115 patients with either biopsy-proven NASH (55; 48%) or cryptogenic cirrhosis thought secondary to NASH (60; 52%) were analyzed for the primary outcome.

Two hundred and forty-one patients (181 Northwestern, 60 University of Chicago) with biopsy-proven ETOH cirrhosis underwent LT during the study period. One hundred and fourteen patients were excluded because of active malignancy (41 patients), other possible causes of liver disease (38 patients), or insufficient data (35 patients). A total of 127 patients with ETOH cirrhosis served as the final comparison group (Fig. 1).

Figure 1.

Flow diagram of study subjects. *Either biopsy-proven NASH (55; 48%) or cryptogenic cirrhosis thought secondary to NASH (60; 52%).

Patients were considered to have a diagnosis of NASH cirrhosis if they had histopathology consistent with NASH, as defined by the modified Brunt criteria,16 and fulfilled the following clinical criteria: absence of alcohol history, careful exclusion of all other forms of chronic liver disease (CLD) based on a combination of history, laboratory testing, histology, and absence of exposure to hepatotoxins or medications associated with hepatic steatosis. All explant biopsy reports and available pretransplant liver biopsy reports were reviewed to confirm the etiology. Patients who did not have a histopathologic diagnosis of NASH were considered to have NASH based on established clinical criteria.17 Because there were no demographic differences (see Table 1) between subjects who had a pretransplant diagnosis of NASH and those with cryptogenic cirrhosis, the study group was termed NASH collectively.

Table 1. Differences in Baseline Characteristics of Patients Undergoing LT for NASH Versus ETOH Cirrhosis
CharacteristicNASH (N = 115)ETOH (N = 127)P Value*
  • Abbreviation: SD, standard deviation.

  • *

    t test for continuous variables; chi-square test or Fischer's exact test for categorical variables.

  • Traditional cardiac risk factors (0-5): males age >45, females age >55, personal history of DM, HTN, or HL and family history of coronary artery disease.

Age, mean, years58.4 ± 9.953.3 ± 9.3<0.001
Female, no. (%)52 (45)23 (18)<0.001
Race/ethnicity, no. (%)   
 White89 (77)83 (65)0.99
 Hispanic19 (17)14 (11)0.99
 African American1 (1)8 (6)0.98
 Other6 (5)00.89
MELD score, mean24.7 (7-40)26.2 (9-40)0.18
Simultaneous liver-kidney transplant, no. (%)32 (28)35 (28)0.99
BMI (kg/m2), mean32.1 ± 7.428.3 ± 6.60.36
 BMI ≥35, no. (%)37 (32)12 (9)<0.01
Number of ATP III criteria met, mean2.3 ± 1.31.1 ± 1.0<0.01
 Waist circumference (males >40 in, females >45 in), no. (%)64 (56)10 (8)0.18
 Hypertriglyceridemia (>150 mg/dL), no. (%)29 (25)8 (6)<0.01
 Low HDL (males <40 mg/dL; females <50 mg/dL), no. (%)48 (42)25 (20)<0.05
 Blood pressure >130/85 mmHg or anti-HTN treatment, no. (%)61 (53)48 (38)<0.05
 DM or fasting glucose >110 mg/dL, no. (%)59 (51)66 (52)0.62
Current or former smoker (within 5 years of transplant), no. (%)34 (30)55 (43)<0.001
Cardiovascular medical history, no. (%)   
Myocardial infarct7 (6)3 (2)0.26
CAD23 (20)11 (9)0.05
HL25 (22)7 (6)<0.01
Congestive heart failure8 (7)8 (6)0.61
Atrial fibrillation7 (6)5 (4)0.50
Stroke9 (8)3 (2)0.12
History of chronic renal insufficiency, no. (%)35 (30)49 (39)0.12
Cardiac risk assessment   
Number of traditional cardiac risk factors, mean2.5 ± 1.32.03 ± 1.1<0.05
Revised cardiac risk index, mean1.79 ± .771.81 ± .770.71
Family history of CAD, no. (%)26 (23)35 (28)<0.05
Systolic blood pressure120.1 ± 20.6121.7 ± 23.90.68
Diastolic blood pressure63.9 ± 11.965.8 ± 13.60.14
Laboratory values, mean ± SD   
Creatinine2.11 ± 1.71.93 ± 1.40.37
Creatinine ≥2.0, no. (%)39 (34)46 (36)0.48
Alanine aminotransferase57.3 ± 70.967.3 ± 123.9<0.05
Total bilirubin8.3 ±11.38.0 ± 7.70.79
Fasting glucose128.1 ± 52.2122 ± 46.10.35
Hemoglobin A1c5.9 ± 1.35.8 ± 0.880.88
Sodium135.2 ± 3.7134.2 ± 7.30.53
Low-density lipoprotein80.3 ± 38.775.9 ± 35.60.56
HDL37.9 ± 21.037.1 ± 19.80.84
TGs108.1 ± 93.979.3 ± 42.9<0.05
Preoperative medication management, no. (%)   
Beta-blocker58 (50)59 (46)0.68
Aspirin13 (11)3 (2)<0.05
Anti-HTNs48 (42)47 (37)0.89
Statin11 (10)1 (1)<0.01

Patients were considered to have a diagnosis of ETOH cirrhosis based on medical history and consistent histopathology.18 In addition, patients underwent careful exclusion of all other forms of CLD, as described above.

All patients received intravenous steroids during, and for a few months after, LT. Maintenance immunosuppression regimens typically included a calcineurin inhibitor (e.g., cyclosporine or tacrolimus) with or without a cell-cycle inhibitor (e.g., mycophenolic acid or sirolimus).

Data Collection and Definitions.

Data collected and analyzed included age, sex, race, ethnicity, body mass index (BMI), waist circumference, past medical, smoking history, and family history, preoperative medication use, laboratory data, Model for End-Stage Liver Disease (MELD) score, infectious complications, cardiovascular complications, retransplantation, and cause of death, as appropriate. Preoperative cardiac assessments were as follows:

  • 1.Electrocardiography (ECG) to evaluate for arrhythmia and QT interval prolongation. Prolonged corrected QT interval (QTc) was defined as ≥440 ms and was calculated using Bazett's formula (QTc = QT/RR1/2).
  • 2.Echocardiography to evaluate ventricular systolic and diastolic contractile performance. Diastolic dysfunction was diagnosed by an increase in E wave deceleration time (EDT) ≥230 msec and a decreased E/A ratio (E velocity/A velocity) <1.1.
  • 3.Dobutamine-stress echocardiography (DSE) or adenosine-thallium myocardial perfusion imaging was used to exclude coronary artery disease (CAD). An abnormal scan was defined as >1 wall motion abnormalities or perfusion defects poststress.
  • 4.Coronary angiography to detect coronary atherosclerotic disease
  • 5.Right heart catheterization to measure mean pulmonary artery pressure

Not all preoperative tests were performed in every patient. Based on a physician's assessment, some patients had no or minimal risk factors for developing cardiac complications, as defined by the American Association for the Study of Liver Diseases practice guidelines.19 Thus, some patients with a “low cardiac risk profile” did not undergo stress testing. The numbers (%) of patients who had these tests available for review were as follows: ECG (n = 238; 98.3%); echocardiography (n = 197; 81.4%); DSE (n = 97; 40.1%); nuclear stress imaging (n = 59; 23.4%); coronary angiography (n = 99; 41.0%); and right heart catheterization (n = 72; 29.8%. Any stress-induced ischemia detected prompted a referral to cardiology and left heart catheterization. If the DSE was nondiagnostic as a result of inability to achieve >85% maximal predicted heart rate, patients were referred to cardiology to direct further workup.

CV complications were defined as the composite outcome of death from any cardiac cause, nonfatal myocardial infarction, new-onset heart failure, cardiac arrest, supraventricular tachycardia requiring intervention, atrial fibrillation or flutter, symptomatic stable ventricular tachycardia requiring treatment, complete heart block, or stroke. Congestive heart failure included findings of either new-onset systolic or diastolic dysfunction or acute cardiogenic pulmonary edema, defined by the presence of dyspnea at rest, tachypnea, tachycardia, severe hypoxemia, brain natiuretic peptide >100, and rales or wheezing on exam. The diagnosis of myocardial infarction was based on the presence of typical ischemic ECG changes and elevated cardiac enzymes. Stroke was defined by the documentation of loss of earlier neurologic function in the setting of imaging consistent with decreased cerebral perfusion. Complications were counted as present or absent only once in the cohort; 1 patient had both a supraventricular arrythmia and a cardiac arrest, 1 patient had a myocardial infarction and congestive heart failure, and 1 patient had both congestive heart failure and a cardiac arrest. Documentation of CV event recorded was based on the first documented event in medical records. Patient survival was defined as time from transplantation to death or last follow-up. Graft survival was defined as time from transplantation to retransplantation or last follow-up.

Metabolic syndrome was defined using the guidelines from the 2004 Revised National Cholesterol Education Program–Adult Treatment Panel (ATP) III.20 Metabolic syndrome was defined as present if patients had more than three 3 of the following: (1) abdominal obesity (waist girth = 102 cm for men; 88 cm for women); (2) triglycerides (TGs) >150 mg/dL; (3) high-density lipoprotein (HDL) <40 mg/dL in men and <50 mg/dL in women; (4) blood pressure >130/85 mm/Hg; and (5) fasting plasma glucose >110 mg/dL.

For the purpose of our study, patients also met criteria if being treated for relevant disease, including dyslipidemia, hyperglycemia, or systemic hypertension (HTN).

Statistical Analysis.

The significance of univariate association of 1-year CV events with pretransplant characteristics was evaluated using a variable reduction strategy using t tests (or rank-sum tests) for continuous variables and chi-square tests for categorical variables. Because both age and sex are known predictors of CV disease, we initially examined the incidence of post-transplant CV events for both males and females using age-adjusted univariate analysis for the primary outcome. There was no significant difference in perioperative CV event rates, thus a combined age- and sex-adjusted model was used. Variables with a value of P ≤ 0.10 from the univariate model were then included in multivariate analyses, including recipient age, sex, smoking status, pretransplant diabetes, CV disease, and the presence of metabolic syndrome. Cox's regression analyses were performed on the following variables for patient and graft survival: recipient age; sex; race; and pretransplant MELD score.

Differences in patient characteristics and postoperative outcomes between NASH and ETOH patients were compared using the Student t test for normal and nonparametric tests for non-normal continuous variables. The chi-square test was used for categorical variables. Kaplan-Meier's survival analysis with the log-rank test and Cox's proportional hazard models were used to compare hazard ratios (HRs) for time to death. Logistic regression models were estimated to compare the likelihood of a CV event within 1 year of transplantation between groups controlling for relevant covariates, including transplant center. Secondary endpoints were graft and patient survival. Analyses were performed using SPSS 19 software (SPSS, Inc., Chicago IL).

Results

Patient Characteristics.

Baseline clinical, demographic, and biochemical characteristics of the two groups are shown in Table 1. The NASH cohort was older (58.4 versus 53.3 years; P < 0.001) and was more likely to be female (45% versus 18%; P < 0.001) than patients who underwent transplantation for ETOH. As expected, the proportion with morbid obesity (32% versus 9%; P < 0.01) was significantly higher in the NASH group. These patients were more likely to meet criteria for metabolic syndrome and therefore had a higher incidence of dyslipidemia and hypertension. NASH patients were also more likely to be on statin or aspirin therapy before transplant. NASH subjects were less likely than ETOH subjects to have a history of smoking (30% versus 43%; P < 0.001) and were less likely to have a family history of CAD (23% versus 28%; P < 0.05). NASH patients, on average, had 2.5 of 5 traditional cardiac risk factors (age: male >45 or female >55 years, personal history of diabetes mellitus [DM], HTN, hyperlipidemia [HL], or family history of CAD), compared to 2 in the ETOH group (P < 0.05). In addition, there was a trend toward a higher incidence of CAD in the NASH group, when compared to ETOH patients (P = 0.05). Of note, there was no difference in preoperative MELD score or the number of patients undergoing simultaneous liver-kidney transplant.

Differences in preoperative cardiac testing are shown in Table 2. NASH patients were more likely to have had a preoperative transthoracic echocardiogram performed (90% versus 76%; P < 0.01). Approximately 60% of all patients underwent noninvasive stress imaging, and 40% underwent cardiac catheterization. There were no significant differences in the number of positive tests before LT, with the exception that ETOH subjects were more likely to have left ventricular hypertrophy on echocardiography (20% versus 15%; P < 0.05) and were more likely to have a negative cardiac catheterization (20% versus 14%; P < 0.05) before LT.

Table 2. Differences in Preoperative Cardiac Testing in Patients Undergoing LT for NASH Versus ETOH Cirrhosis
CharacteristicNASH (N = 115)ETOH (N = 127)P Value*
  • Abbreviation: EF, ejection fraction.

  • *

    t test for continuous variables; chi-square test or Fischer's exact test for categorical variables.

Transthoracic echocardiogram performed, no. (%)103 (90)94 (74)<0.01
 EF <55% on transthoracic echocardiogram4 (9)9 (7)0.08
 Left ventricular hypertrophy present17 (15)26 (20)<0.05
 Pulmonary arterial systolic pressure ≥50 mmHg7 (6)3 (2)0.16
 Diastolic dysfunction present17 (15)20 (16)0.34
 Moderate or severe valvular disease19 (17)17 (13)0.95
Noninvasive stress imaging performed74 (64)71 (56)0.37
 DSE performed50 (43)47 (37)0.57
  ≤85% of maximal predicted heart rate achieved21 (18)24 (19)0.19
 Myocardial perfusion imaging performed29 (25)32 (25)0.84
  Ischemia noted on noninvasive stress imaging9 (8)4 (3)0.35
Left heart catheterization performed51 (44)48 (38)0.16
 No CAD16 (14)25 (20)<0.05
 Nonobstructive CAD31 (27)24 (19)0.50
 Obstructive CAD7 (6)5 (4)0.90
Right heart catheterization performed36 (31)36 (28)0.09
 Mean pulmonary arterial pressure ≥25 mmHg15 (13)19 (15)0.72
QTc on ECG ≥440, no. (%)74 (64)78 (61)0.81
Coronary artery calcifications seen on noninvasive imaging25 (22)3 (2)0.83

Cardiovascular Complications.

On both uni- and multivariate analysis, NASH patients were more likely to experience an adverse CV event <1 year after LT, compared to ETOH patients, even after controlling for age, sex, BMI, smoking, previous history of metabolic syndrome and previous history of CAD (odds ratio [OR] = 4.12; 95% confidence interval [CI] = 1.91-8.90). In the NASH group, 26.4% of patients had an adverse CV event within 1 year of LT versus 8.2% in the ETOH group (P < 0.001). The most common cardiac complication in both groups was acute pulmonary edema (18.1% NASH versus 16.2% ETOH; P = 0.33), followed by new-onset atrial fibrillation (10.3% NASH versus 8.4% ETOH; P = 0.61). NASH patients were significantly more likely to experience a sudden cardiac arrest (8.3% versus 1.3%; OR = 5.37; 95% CI: 1.13-25.4), compared to ETOH patients. There was no difference in the incidence of myocardial infarction, stable ventricular tachycardia, supraventricular tachycardia, new-onset heart failure, or stroke (Table 3).

Table 3. CV Events in Patients Transplanted for NASH Versus ETOH Cirrhosis
CharacteristicNASH (N = 115)ETOH (N = 127)NASH Versus ETOH, OR* (95% CI)P Value
  • *

    Adjusted for recipient age, sex, smoking status, pretransplant diabetes, CV disease, and the presence of metabolic syndrome.

  • t test for continuous variables; chi-square test or logistic regression for categorical variables.

  • Some patients had more than one cardiac complication; only the first event is counted here.

Any CV event within 1 year of transplant, no. (%)30 (26)10 (8)4.12 (1.91-8.90)<0.001
 Acute pulmonary edema21 (18)30 (16)0.73 (0.39-1.37)0.33
 New-onset atrial fibrillation11 (10)10 (8)1.26 (0.52-3.09)0.61
 Cardiac arrest9 (8)2 (1)5.37 (1.13-25.39)<0.05
 Acute heart failure3 (3)10 (8)0.31 (0.08-1.16)0.07
 Stroke6 (5)7 (6)0.95 (0.31-2.90)0.92
 Stable ventricular tachycardia2 (2)0 (0)1.02 (0.99-1.04)0.14
 Supraventricular tachycardia2 (2)1 (1)2.23 (0.20-24.98)0.92
 Non-ST elevation myocardial infarction2 (2)3 (2)0.73 (0.12-4.47)0.74
 ST elevation myocardial infarction1 (1)2 (1)0.73 (0.12-4.47)0.74

Table 4 shows characteristics of the 30 NASH subjects who had at least one cardiovascular event within the first year of LT. The majority (70%) of the events occurred in the immediate perioperative period, and 90% of patients who had an event had undergone previous noninvasive stress imaging. However, 37% of these patients did not achieve maximal heart rate. Of the 11 patients with a suboptimal stress test, 6 (54%) proceeded to cardiac catheterization and 3 (45%) then underwent myocardial perfusion imaging. The 2 patients who proceeded to transplant in the setting of a suboptimal preoperative stress test developed acute heart failure and pulmonary edema within the first 30 days after transplant. Forty percent of NASH patients with a CV event had an invasive coronary angiography performed, and a significant proportion (58%) of those patients had no or minimal CAD. Two of these patients went on to have an acute myocardial infarction, 2 had a subsequent stroke, 1 had a cardiac arrest, and the remaining 2 patients developed acute heart failure post-transplant. Finally, 77% of NASH patients who had a cardiac event were noted to have had a prolonged QT interval on preoperative ECG.5

Table 4. Characteristics of 30 NASH Subjects With at Least One CV Event Within 1 Year of LT
CharacteristicNumber of Subjects (%)
Timing, days 
 Intraoperative2 (7)
 Perioperative (0-30 days)19 (63)
 Early postoperative (1-6 months)7 (23)
 Late postoperative (6 months to 1 year)1 (3)
Pretransplant cardiac evaluation, no. (%) 
 Transthoracic echocardiogram performed30 (100)
 Noninvasive stress imaging performed27 (90)
  ≤85% of maximal predicted heart rate achieved11 (37)
 Left heart catheterization performed12 (40)
  No CAD2 (7)
  Nonobstructive CAD5 (17)
  Obstructive CAD5 (17)
 Right heart catheterization performed8 (27)
 QTc on ECG ≥44023 (77)
 Coronary artery calcifications observed on noninvasive imaging9 (30)
Preoperative medication management, no. (%) 
 Beta-blocker17 (57)
 Anti-HTNs15 (50)
 Aspirin5 (17)
 Statin4 (13)
Preoperative CV risk factors, no. (%) 
 Central obesity (BMI >25 kg/m2)16 (53)
 Hypertriglyceridemia (>150 mg/dL)8 (27)
 Low HDL (males <40 mg/dL; females <50 mg/dL)13 (43)
 Elevated low-density lipoprotein (>160 mg/dL)7 (23)
 HTN18 (60)
 DM or fasting glucose >110 mg/dL17 (57)
  DM on insulin9 (30)
 History of chronic renal insufficiency11 (37)
 Family history of CAD12 (40)
 Current or former smoker (within 5 years of transplant)11 (37)
 Personal history of CAD4 (13)
Overall patient survival, no. (%)15 (50)
 Death from CV event8 (27)
 Death from sepsis6 (20)
 Death from other cause4 (13)
Table 5. Comparison of Post-Transplant Outcomes in NASH Versus ETOH Patients
CharacteristicNASH (N = 115)ETOH (N = 127)NASH Versus ETOH, OR (95% CI)P Value*
  • Abbreviation: SD, standard deviation.

  • *

    t test for continuous variables; chi-square test or logistic regression for categorical variables.

  • **Other causes: malignancy (9); graft failure (2); hemorrhage (4); respiratory failure (3); and unknown (6).

Infectious complications, no. (%)    
 Sepsis within 30 days of transplant27 (17)17 (6)1.94 (0.99-3.79)0.05
 Infection within 1 year of transplant requiring hospitalization43 (36)41 (30)1.23 (0.72-2.10)0.45
Length of stay, mean days ± SD11.20 ± 11.4011.0 ± 12.20 0.15
 Intensive care unit stay, mean days6.36 ± 11.414.52 ± 6.21 0.86
Overall patient survival, %72.374.61.12 (0.63-2.00)0.69
 At 1 year81.388.11.71 (0.83-3.50)0.14
 At 3 years73.385.32.07 (1.02-4.20)<0.05
 At 5 years60.368.81.45 (0.75-2.82)0.27
Cause of death, no. (%)    
 Death from sepsis12 (11)6 (4)2.26 (0.82-6.24)0.11
 Death from CV event10 (9)4 (1)2.72 (0.83-8.95)0.09
 Death from other cause10 (9)14 (11)0.70 (0.30-1.65)0.41
Overall graft survival, %78.698.72.29 (0.99-5.08)0.06
Follow-up time, mean months ± SD50.84 ± 36.0559.97 ± 40.76 0.73

Over 50% of the NASH patients who had a CV event had underlying risk factors for CV disease and metabolic syndrome, most commonly dyslipidemia or HTN (Table 4). However, preoperative medication management for these conditions was suboptimal; only 13% of patients received a statin and only 17% an aspirin. Beta-blocker use was more common, yet 43% of patients who had a subsequent CV event were not receiving preoperative beta-blockade. Experiencing a CV event in the NASH cohort was associated with a 50% overall mortality, with 27% of deaths caused by the CV event itself and 20% caused by a primary diagnosis of sepsis.

Survival Analysis.

Despite poorer survival probabilities for NASH patients, differences were not significant between the NASH and ETOH cohorts (HR = 0.82; 95% CI: 0.49-1.34; Fig. 2). The 1-, 3-, and 5-year patient survival probabilities were 81.3%, 73.3%, and 60.3% in the NASH group and 88.1%, 85.3%, and 68.8% in the ETOH group.

Figure 2.

Kaplan-Meier curves comparing survival rate for patients transplanted for NASH versus ETOH-induced cirrhosis.

Sepsis was the most common cause of death post-LT (11% versus 4% in the ETOH group; P = 0.11), followed by CV causes in the NASH group (9% versus 1% in the ETOH group; P = 0.13). There were no differences in the rates of early sepsis (<30 days after transplant) or in the total number of infectious complications requiring hospitalization within 1 year of transplant (36% NASH versus 30% ETOH; p = 0.45). In the ETOH group, death was most commonly related to sepsis, graft failure, or malignancy after transplant. There was also no difference in graft survival (HR = 0.93; 95% CI: 0.71-1.23; P = 0.21; Fig. 3), even when controlled for recipient age, sex, race, and pretransplant MELD score.

Figure 3.

Kaplan-Meier curves comparing graft survival rate for patients transplanted for NASH versus ETOH-induced cirrhosis.

Discussion

Our results demonstrate that NASH patients have an increased risk of post-transplant adverse CV events, when compared to patients transplanted for alcohol-induced liver disease, even after controlling for traditional preoperative cardiac risk factors.

A growing body of research now recognizes an entity known as “cirrhotic cardiomyopathy”.2, 3 Although the clinical presentation can be variable, all patients have four common features: (1) baseline increased cardiac output; (2) attenuated systolic contraction and diastolic relaxation; (3) electrophysiological abnormalities, including repolarization change; and (4) a reduced response of the heart to direct beta stimulation (beta-incompetence).2 These changes occur in the absence of overt congestive failure. In our cohort, arrhythmia and cardiac arrest accounted for over 50% of all cardiac complications, suggesting that this physiologic phenomenon may be partly responsible for the high rate of adverse CV events observed in both groups.

Interestingly, 77% of NASH patients who had an adverse CV event were also noted to have a prolonged QTc interval on preoperative ECG. Abnormalities in cardiac electrophysiology are well documented in patients with cirrhosis, and the prolonged QTc interval has emerged as the electrophysiological hallmark of cirrhotic cardiomyopathy.21 In patients with cirrhosis, the mechanisms of QTc prolongation remain unclear; however, it is associated with electrolyte abnormalities, myocardial ischemia, and sympathetic hyperactivity.22-24 During LT, these factors can be further aggravated by hypovolemia resulting from bleeding and compression of the inferior vena cava during the anhepatic phase.25 Additionally, various drugs affecting the QTc interval are administered before and during LT.26, 27 Finally, insulin resistance and the presence of metabolic syndrome are independent risk factors for QTc prolongation, which may further increase CV morbidity and mortality in NASH patients who undergo LT.28, 29

Consistent with previous research, our study demonstrates that NASH patients who undergo LT have overall comparable survival to other etiologies.13, 14, 30 Although a higher number of patients died from CV causes in the NASH group (9%), compared to the ETOH group (1%), the difference did not reach statistical significance (P = 0.13). However, CV mortality in our NASH group was higher than that based on published autopsy results for all LT patients.31 In addition, Albedawi et al. recently demonstrated that CV disease risk after LT varies with etiology of liver disease and is higher for NASH cirrhosis than in comparison to those patients transplanted for primary biliary cirrhosis or primary sclerosing cholangitis.32 These results suggest the need for an individualized approach to cardiac assessment and risk stratification pretransplant that includes consideration for not only traditional CV risk factors, but also etiology of liver disease, especially in patients with NASH cirrhosis.

Despite a high preoperative utilization of invasive and noninvasive CV imaging, CV events were still common overall. For instance, DSE, which is routinely used at our institution as a means of preoperative risk stratification, was suboptimal in 20% of patients who failed to achieve the target heart rate, which decreases its predictive ability.33 Furthermore, cardiac catheterization was performed in a significant proportion of patients (>40%); however, significant stenosis requiring intervention was found in <10%. Even myocardial perfusion defects with a normal angiogram are not necessarily benign findings and could be caused by reduced microvascular coronary blood flow.34 Questions therefore remain as to what degree of stenosis requires intervention and what type of treatment is best. Most subcritical coronary stenoses in surgical patients can be medically managed. However, investigators report a perioperative death rate of more than 50% in transplant recipients with CAD who were managed medically.35 Further myocardial infarction during transplant surgery has occurred with as little as 30% vessel occlusion.36 Our study supports the growing body of evidence showing that current CV predictive tests used in other surgical realms may underperform in their predictive value in LT patients. This underscores the need for large, prospective studies examining CV risk stratification in LT patients.

Finally, in our study, we note that a minority of patients were receiving pretransplant medical management for their cardiovascular comorbidities. Several studies have demonstrated reductions in perioperative cardiac events and mortality when high-risk patients take beta-blockers before noncardiac surgery.37-39 Recently, it has been reported that use of perioperative beta-blockers is protective (P = 0.004; OR, 0.20; 95% CI, 0.1-0.6) for combined cardiac outcomes in an LT population.40 Yet, only 50% of patients in both of our cohorts were on beta-blocker therapy before transplant, and only 57% of NASH patients who experienced a CV event were receiving preoperative beta-blockade that may have had an indication for medication use. Because of the retrospective nature of the study, reasons for medication underuse, such as severe portal hypertension leading to hypotension, are difficult to ascertain.

Statins are frequently used to reduce CV mortality and morbidity, and statins may be of benefit in the perioperative period.41 Statins clearly reduce CV morbidity, irrespective of the patients' cholesterol status in elective and emergency surgery.42 Beyond cholesterol reduction, statins have multiple beneficial influences on vascular endothelial function, atherosclerotic plaque stability, inflammation, and thrombosis.43, 44 In fact, the pleiotropic actions of statins on endothelial function have emerged as new therapy to reduce portal pressure by targeting multiple molecular pathways involved in hepatic vascular homeostasis.45 Many studies have demonstrated the safety and benefit of statin use in patients with advanced liver disease.46, 47 Using the ATP III guidelines20 for recommended statin use, only 11 of 31 (35.5%) NASH patients and 1 of 14 (7.1%) ETOH patients with a clear indication for statin use were receiving statin therapy at the time of transplant in our study. These findings emphasize the need for further prospective studies to evaluate the potential benefit of intensified perioperative medication management in LT candidates, specifically those at increased risk of adverse CV events undergoing transplant for NASH cirrhosis.

There are limitations of our analysis. Because of the study's retrospective nature, it was not feasible to obtain all baseline characteristics. In addition, a large number of patients (66 patients) were excluded for insufficient data (Fig. 1). The decision to pursue preoperative cardiac testing was determined by the treating physician, leading to several possible sources of bias in our data. The strength of this study, however, is that it examines only biopsy-proven or explant histology demonstrating a diagnosis of NASH cirrhosis. The exclusion of all other patients with presumed NASH led to smaller numbers in our cohort.

In summary, our study demonstrates that NASH patients have an increased risk for the development of CV complications within 1 year after LT, compared to ETOH patients, even after controlling for preexisting comorbidities and despite extensive preoperative testing. CV complications do not appear to alter the overall survival of patients who undergo LT for cirrhosis resulting from NASH, compared to ETOH. This study also identifies a deficiency in our current medical practice to adequately address the metabolic comorbidities in NASH patients before transplant. Devising prospective studies to address CV risk stratification and aggressive management of metabolic syndrome before transplant may lead to real changes in the natural history of LT resulting from NASH cirrhosis.

Ancillary