- Top of page
- Patients and Methods
- Supporting Information
The survival of patients with end-stage liver disease (ESLD) after liver transplantation (LT) is highly dependent on cardiac function. A significant proportion of patients with ESLD have cardiac dysfunction, including high output cardiac failure, valvular heart disease and pulmonary venous/arterial hypertension, all of which affect peri-operative survival [1, 2]. Previous studies have shown that hemodynamic parameters from right heart catheterization (RHC), that is, mean pulmonary artery pressure (MPAP) and pulmonary vascular resistance (PVR), are predictive of outcomes after LT [2, 3]. In these studies, MPAP ≥ 50 mmHg was associated with 100% mortality in the peri-operative period, whereas MPAP between 35–50 mmHg correlated with increased mortality in the subset of patients with PVR ≥ 250 dynes s/cm5. Most of the deaths in this cohort were cardiac in nature (right heart failure, ventricular arrhythmias), suggesting a higher rate of cardiac mortality in patients with unfavorable pre-LT hemodynamics . While RHC provides useful predictive hemodynamic data, it is invasive, relies on single, “snap-shot” measurements, and does not provide an assessment of myocardial structure and function.
Two-dimensional (2D)/Doppler echocardiography has been widely used as a screening modality for estimating cardiac hemodynamics and the need for RHC [4, 5]. Echocardiography has become standard of care in the evaluation of patients undergoing LT and is helpful for excluding portopulmonary hypertension . Despite its use as a screening modality for cardiac abnormalities, portopulmonary hypertension and the need for RHC, echocardiographic evaluation of the right heart has not yet been studied as an independent predictive tool for post–operative outcomes.
Therefore, we sought to determine whether echocardiographic indices of right heart structure and function were independently associated with morbidity and mortality post-LT. We hypothesized that abnormalities in right heart structure and function correlated with adverse outcomes post-LT.
- Top of page
- Patients and Methods
- Supporting Information
Of the 291 patients receiving LT during the study period, 257 patients underwent pretransplant 2D/Doppler echocardiography with images available for review. From this subset, 7 HIV positive patients, 8 re-LT patients and 26 patients with incomplete Doppler images were excluded. A total of 216 patients met the inclusion and exclusion criteria and underwent echocardiography at a median 71 (IQR: 15–160) days prior to LT (Table 1). The mean age of the cohort was 55 ± 11 years and patients were predominantly male (62%). The majority of patients were white (72%), and the prevailing etiology of liver disease was hepatitis C (37%), followed by alcohol (30%), nonalcoholic steatohepatitis/cryptogenic (18%) and other (15%). The average MELD score at the time of transplantation was 25 ± 11 and the mean post-LT follow-up was 598 ± 330 days. No patients were lost to follow-up. Three patients were on therapy for portopulmonary hypertension with pulmonary vasodilators and showed improvement in their clinical and echocardiographic parameters, including PASP and TR prior to LT. At the time of LT, severity of TR in this group was ≤mild TR, and did not change post-LT. A summary of echocardiographic measurements can be found in Table 2.
Table 1. Patient characteristics
|Patient characteristics (n = 216)|
|Age—years||56 ± 11|
|Male sex—no. (%)||133 (62)|
|African American||11 (5)|
|Etiology of ESLD—no. (%)|
|Hepatitis C||80 (37)|
|Medical comorbidities—no. (%)|
|Coronary artery disease||20 (9)|
|Diabetes mellitus||51 (24)|
|Chronic obstructive pulmonary disease||11 (5)|
|Chronic kidney disease (non–hepatorenal syndrome)||8 (4)|
|Pre-LT renal replacement therapy||50 (23)|
|Complications of liver disease—no. (%)|
|Hepatic encephalopathy||103 (48)|
|Spontaneous bacterial peritonitis||34 (16)|
|Hepatorenal syndrome||50 (23)|
|Esophageal varices||102 (47)|
|Transjugular intrahepatic portosystemic shunt||21 (10)|
|Hepatocellular carcinoma||54 (25)|
|Labs at time of LT|
|International normalized ratio||2.0 ± 1.2|
|Creatinine (mg/dL)||1.8 ± 1.5|
|Total bilirubin (mg/dL)||10.8 ± 12.9|
|Alanine aminotransferase (unit/dL)||212 ± 816|
|Aspartate aminotransferase (unit/dL)||311 ± 1284|
|Sodium (mEq/L)||136 ± 5|
|MELD score at time of LT||25 ± 11|
|Mean post-LT follow-up—days||598 ± 330|
Table 2. Echocardiographic measurements
|Echocardiographic characteristic||Mean and standard deviation (except where noted differently)||Reference standard per American Society of Echocardiography Guidelines|
|Septal wall thickness (cm)||1.0 ± 0.2||0.6–1.0|
|Posterior wall thickness (cm)||1.0 ± 0.2||0.6–1.0|
|LV end-diastolic dimension (cm)||4.6 ± 0.7||3.9–5.9|
|LV end-systolic dimension (cm)||2.7 ± 0.6||2.0–4.0|
|Left atrial dimension (cm)||3.6 ± 0.7||2.7–4.0|
|LV mass index (g/m2)||80.5 ± 20.9||44–88 (varies by male/female)|
|LV ejection fraction (%)||63.3 ± 5.6||>55|
|RV diameter, parasternal long-axis (cm)||3.7 ± 0.6||—|
|RV diameter, parasternal short-axis (cm)||3.6 ± 0.6||—|
|RV maximal diameter, apical 4-chamber view (cm)||4.0 ± 0.7||2.4–4.2|
|RV end-diastolic area (cm2)||23.7 ± 6.5||11–28|
|RV end-systolic area (cm2)||14.8 ± 6.6||4–15|
|RV fractional area change||0.4 ± 0.1||0.35–0.63|
|Tricuspid annular plane systolic excursion (cm)||2.4 ± 0.8||1.6–3.0|
|RV/LV maximal diameter ratio (measures RV dilatation)||0.8 ± 0.1||<0.6 normal; 0.7–0.9 moderate; >1 severe|
|Tricuspid regurgitation (presence, severity)—no. (%)|
|Pulmonary artery systolic pressure (mmHg)||38.21 ± 11.57||—|
|Stroke volume (mL)||84.98 ± 26.42||—|
|Cardiac output (L/min/m2)||9.76 ± 2.43||—|
On univariate analysis, PASP, severity of TR (≥mild TR), pre-LT RRT, pre-LT SBP and MELD score at LT was found to be significantly associated with adverse outcomes, for example patient and graft survival, ICU LOS, total LOS, time on ventilator, need for post-LT RRT and post–operative infection (tables in Supplementary Section). Other measures of right ventricular structure and function, including RV FAC, TAPSE and RV/LV ratio, were not statistically significant on univariate analysis.
Of the variables that were significant (p ≤ 0.15) on univariate analysis, only severity of TR was associated with both patient and graft survival on multivariate analysis (Table 3). This relationship is further demonstrated in the Kaplan–Meier curves (Figures 1 and 2) for patient and graft survival according to the severity of TR (p = 0.002 and p < 0.001, respectively). Regarding predictors of other post-LT short-term outcomes (Table 4), MELD at LT, pre-LT RRT and pre-LT PASP on echocardiography were associated with total LOS, time on ventilator and the need for post-LT RRT. MELD and PASP were additionally associated with ICU LOS and MELD alone with post-LT infection. Neither severity of TR nor history of SBP was predictive of these short-term outcomes.
Table 3. Multivariate predictors of patient and graft survival
|Pre-LT variable||Patient survival||Graft survival|
|Hazard ratio||95% Confidence interval||p-Value||Hazard ratio||95% Confidence Interval||p-Value|
|MELD score at time of LT||1.04||0.97–1.09||0.07||1.03||0.99–1.07||0.18|
|Severity of TR (≥mild TR)||3.91||1.62–9.44||0.002||3.70||1.70–8.06||0.001|
Figure 1. Patient survival following liver transplantation based on the presence or absence of ≥mild tricuspid regurgitation (log rank test, p = 0.0018), where 0 = none or trace TR and 1 = mild, moderate or severe TR.
Download figure to PowerPoint
Figure 2. Graft survival following liver transplantation based on the presence or absence of ≥mild tricuspid regurgitation (log rank test, p = 0.0006), where 0 = none or trace TR and 1 = mild, moderate or severe TR.
Download figure to PowerPoint
Table 4. Multivariate predictors of short-term post-LT outcomes
|Pre-LT variable||ICU stay (days)||LOS (days)||Ventilator time (days)||Post-LT RRT||Infection|
|CE||SE||p-Value||CE||SE||p-Value||CE||SE||p-Value||OR||95% CI||p-value||OR||95% CI||p-value|
|MELD score at time of LT||0.22||0.08||0.005||0.39||0.10||<0.001||0.17||0.06||0.004||1.09||1.04–1.15||<0.001||1.07||1.02–1.13||0.01|
|Severity of TR (≥mild TR)||1.09||1.42||0.45||—||—||—||—||—||—||—||—||—||—||—||—|
Analysis of RV end diastolic area as an additional covariate in the multivariate models did not change any of the associations. Severity of TR remained independently associated with death (p = 0.002) and graft failure (p = 0.001). Furthermore, entering donor characteristics, including donor age, cold ischemia and donor type (i.e. donation after cardiac death) in the multivariate model did not change any of the associations either. Severity of TR remained independently associated with death (p = 0.002) and graft failure (p = 0.001). Finally, analysis of LV systolic and diastolic parameters by TR group (TR none/trace vs. TR mild or greater) showed no significant differences (see table in Supplementary Section).
Overall patient survival at 30 days was 97.7%, at 6 months 93.5%, at 1 year 90.3% and at last follow-up 88.4%. Of the patients that died (N = 25), the predominant cause of death was sepsis (Table 5). Cardiovascular deaths accounted for 20% of deaths, most of which were beyond the 30-day postoperative period. Patients with ≥mild TR accounted for 71% of graft failures and 58% of deaths. The predominant causes of death in the cohort with ≥mild TR was unknown (33%, N = 6) followed by cardiovascular (28%, N = 5) and sepsis (22%, N = 4).
Table 5. Survival data and cause of death
|Characteristic||N = 216|
|Graft survival at 1 year—no. (%)||185 (86)|
|Overall patient survival—no. (%)|
|At 3 months||211 (98)|
|At 6 months||202 (94)|
|At 1 year||195 (90)|
|Primary cause of death in all patients—no. (%)||N = 25|
|Primary cause of death in patients with ≥mild TR—no. (%)||N = 18|
|Timing of patient deaths with ≥mild TR—no. dead (%)||N = 18|
|At 1 month||5 (28)|
|At 3 months||3 (17)|
|At 6 months||2 (11)|
|At 1 year||4 (22)|
|>1 year||4 (22)|
|Primary cause of Re-LT in all patients—no. (%)||N = 6|
|Hepatic artery thrombosis and/or ischemic cholangiopathy||5 (67)|
|Primary nonfunction||1 (17)|
Thirty-one patients had post-LT echocardiograms at a median 8 (IQR: 2–105) days from LT. Seventy-four percent (23/31) had graft failure and echocardiography performed to assess cardiac function post-LT or prior to considering re–transplantation. When comparing pre- and post-LT echocardiograms in this small subset, the severity of TR was the same in 69.6% of patients, worse in 21.7% of patients and improved in 8.7%.
All transplanted patients underwent a liver ultrasound with Doppler flow on postoperative day 1 to assess the liver graft. No evidence of hepatic vascular compromise (i.e. reversal of flow in the hepatic/portal veins or inferior vena cava) was noted in the subset of patients with ≥mild TR.
- Top of page
- Patients and Methods
- Supporting Information
Echocardiography has long been validated as a screening modality for left ventricular dysfunction and pulmonary hypertension in patients undergoing LT [5, 10, 11]. Our data suggest that beyond its utility as a screening tool for left ventricular dysfunction and pulmonary hypertension, pretransplant echocardiographic evaluation of the right heart can be useful in detecting patients at higher risk for significant post-LT outcomes. Specifically, on multivariate analysis, the severity of TR on pre-LT echocardiography was independently predictive of graft and patient survival. In addition, PASP estimation on echocardiography prior to LT was associated with peri-operative outcomes (total and ICU LOS, ventilator time, need for post-LT RRT), as well as other known predictors such as MELD and pre-LT RRT.
Our results have potential implications for transplant candidates and recipients. The severity of TR correlated with 1-year patient and graft survival, which are major initial benchmarks for success with LT. The assessment of right heart structure/function and TR severity is not obtainable from pre-LT RHC, which currently serves as the gold standard in the evaluation of pre-LT cardiac hemodynamics. Our findings suggest that evaluation of severity of TR and PASP may be important in identifying LT recipients at highest risk for complications and should be considered in concert with data obtained from RHC.
The correlation between TR severity and post-LT survival is a novel finding. Only one prior report found a cut-off TR jet velocity of 3.0 m/s was associated with 1-year post-LT survival . However, this study only analyzed TR jet velocity and did not analyze TR severity and other right heart parameters based on ASE guidelines . This finding begs the question as to why TR has such an impact on post-LT graft and patient survival. The pathogenesis may be speculatively related to the backpressure exerted by prolonged TR, leading to eventual hepatic congestion, graft failure and other long-term complications. Alternatively, TR may be a marker of poor reserve or perfusion, highlighted when patients have clinical deterioration (i.e. cardiovascular failure or sepsis). Interestingly, evaluation of post-LT echocardiograms showed that the severity of TR remained similar post-LT in the majority of patients with graft failure, although this was a select group who had this performed post-LT for different reasons and not by protocol. Prospective studies of patients with pre-LT TR are needed to monitor the progression or severity of TR post–operatively, and further studies are needed to determine if the impact of TR pre-LT can be minimized to improve outcomes.
In non–hepatic transplant recipients and non–transplant surgical/medical patients, limited data exist regarding an independent association between TR and morbidity and mortality. The liver graft may be more susceptible to right heart dysfunction than in other medical and surgical settings, given the high requirement for adequate hepatic in-flow and out-flow for optimal function. This is evidenced by the poor outcomes of patients with ESLD who have significant pulmonary hypertension, although deaths in this situation are mainly related to right heart, not allograft, failure [1-3]. In addition, the liver graft at risk for preservation injury, rejection, recurrent viral disease and other types of injury may be even more sensitive to diminished right ventricular function.
The association of pre-LT RRT, PASP on echocardiography and MELD score with adverse short-term postoperative outcomes has been reported previously [2, 3, 12-14]. Specifically, the presence of elevated PASP on echocardiography was shown to have an effect on short-term post-LT outcomes [2, 3]. In our study, however, patients with severe, untreated portopulmonary hypertension were excluded from transplantation and thus not analyzed in this post-LT outcome study. These patients typically fare poorly with transplantation, and in the same vein, our group with elevated PASP on echocardiography (albeit not too high to exclude transplantation) had more immediate complications, possibly as an indication of right heart volume and/or pressure overload in the peri-operative period [1, 2, 15]. An elevated creatinine, a significant component of the MELD score calculation and the need for RRT are well known as risk factors for requiring continued RRT and other associated short-term adverse outcomes [16, 17].
The study has some limitations as a retrospective analysis. Patients with severe right ventricular dysfunction, markedly elevated pulmonary pressures and severe TR on initial cardiac assessment were excluded from the analysis, as these patients were not deemed transplant candidates. Thus, only a small percentage of our study cohort (3.7%) had a subsequent RHC performed pre-LT to confirm intracardiac and pulmonary pressures. This precluded our ability to assess an association between echocardiographic and RHC indices (e.g. MPAP, PVR) as well as their independent or combined prediction of post-LT outcomes. Echocardiographic estimation of PASP, based on several assumptions, including the simplified Bernoulli equation and noninvasive estimation of RA pressure, could have attenuated the associations between PA pressure and outcomes examined in our study. Echocardiographic measurements and images were obtained by different sonographers but separately analyzed by a single trained reader for the purposes of this retrospective study. In a prospective study, a specified protocol with dedicated RV views could provide better images for data extraction, and RAP could be estimated rather than uniformly assigned to be 10 mmHg . In addition, we only used a single time point (one echocardiogram) closest to the LT procedure to correlate with outcomes. Serial echocardiographic measures and, in particular, immediate pre- or intra-operative assessments might have even been more predictive of short- and long-term outcomes. Lack of liver biopsy data was another limitation, which would have allowed for assessment of congestive hepatopathy. Finally, only a small percentage of our patients underwent post-LT echocardiograms, as it was not in our protocol to monitor patients unless clinically indicated. Thus, it is unclear if poor outcomes were associated with post-LT progression of TR or other untested risk factors.
In conclusion, echocardiographic assessments of the right heart may be predictive of adverse post-LT outcomes and useful in the selection and monitoring of LT candidates and recipients. Such data could be used in adjunct to findings from RHC to further risk stratify patients prior to LT. Further prospective studies tied to hemodynamic assessments before and after LT are needed to validate our results.