Patients and Methods
Study Sample
We retrospectively studied consecutive adult (age > 18 years) ESLD patients who underwent LT from 2007 to 2010. At our institution, all LT candidates routinely underwent screening echocardiography. Patients with evidence of right ventricular (RV) systolic dysfunction, septal flattening or severe TR on echo with moderate–severe portopulmonary hypertension (MPAP > 40 mmHg and PVR > 240 dynes s/cm) on RHC that did not respond to vasodilator treatment were not considered candidates for LT at our center and not analyzed as we only included patients who underwent LT. In the LT population, patients with human immunodeficiency virus (HIV) (due to possible underlying HIV associated pulmonary hypertension), those with incomplete pre-LT echocardiographic images and patients undergoing re–transplantation were also excluded from analysis. Follow-up was achieved for at least 1 year in all surviving patients. Pretransplant demographic, clinical and laboratory data were collected for each patient. Data on post-LT outcomes, including patient and graft survival, causes of death, cardiovascular events, total hospital length of stay (LOS), intensive care unit (ICU) LOS, time on ventilator, need for post–operative renal replacement therapy (RRT) and infectious complications, were recorded from the electronic chart and transplant database. Presence of hepatic vascular compromise was evaluated for by reviewing post-LT day 1 liver ultrasounds. Patient survival was defined as the time from transplantation to death or last follow-up, and graft survival was defined as time from transplantation to death or re–transplantation. Patients were censored to time of death or date of last contact. This study was approved by the Northwestern University Institutional Review Board.
Echocardiography
The pretransplant echocardiogram of each patient was analyzed by a single trained reader (L.K.) with oversight from a board certified cardiologist (S.J.S.), blinded to all other data, including outcomes. All right heart measurements were based on criteria outlined by the American Society of Echocardiography (ASE) guidelines for the evaluation of the RV and valvular regurgitation. In cases where patients underwent more than one pre-LT echocardiogram, the one closest to the time of transplantation was analyzed. Patients who underwent post-LT echocardiography were also identified, and their echocardiograms and clinical outcomes were analyzed using the same protocol outlined above. The full details of the standard protocol for echocardiographic assessment at our institution are available in the Supplementary Material.
Multiple echocardiographic measurements were recorded, with specific emphasis on right heart indices: pulmonary artery systolic pressure (PASP), presence and severity of tricuspid regurgitation (TR) and measurements of RV structure and function. The tricuspid regurgitation gradient (TRPG) was calculated from the peak TR velocity using the modified Bernoulli equation: TRPG (mmHg) = 4 V, where V denotes the highest continuous wave Doppler measurement of the TR velocity (m/s) in the parasternal short axis (aortic level), RV inflow and apical 4-chamber views. The right ventricular systolic pressure (RVSP) was estimated by adding the right atrial pressure (RAP) to the TRPG. RAP was uniformly assigned as 10 mmHg, since it could not be estimated from the echocardiograms due to the lack of subcostal views, which precluded measurements of the inferior vena cava. PASP was defined to be equal to RVSP in the absence of RV outflow obstruction or pulmonary valve stenosis. The presence and severity of TR (graded as none/trace, mild, moderate or severe) was qualitatively assessed by evaluating the size, direction and characteristics of the TR jet using Doppler color flow imaging as outlined by the guidelines. Right ventricular diameter, size and area were measured at end diastole in the apical 4-chamber view. Fractional area change (FAC) was calculated in the standard manner [100 × (RV end diastolic area) − (RV end systolic area)/(RV end diastolic area)]. Tricuspid annular plane systolic excursion (TAPSE) was measured in the apical 4-chamber view, as described previously. Finally, data on left heart structure and function, including left atrial dimension, left ventricular end diastolic dimension and septal wall thickness were also collected.
Statistical Analysis
Prior to final statistical analysis, preliminary data were analyzed using univariate and graphical methods wherever applicable, to facilitate inspection and interpretation of the data. Outliers and influential observations were identified and checked for accuracy. All data were summarized using appropriate descriptive statistics (e.g. mean and standard deviation for continuous variables, count and frequency for categorical variables). Kaplan–Meier analysis was used to evaluate patient survival between the TR and non-TR groups using log-rank test.
Data were analyzed using both univariable and multivariable analysis. Linear regression was conducted for continuous outcomes (hospital LOS, ICU LOS and total time on ventilator) while logistic regression was used for binary outcomes (post-LT RRT and infection). Hazard ratios (HR) were calculated using Cox proportional hazards regression model for survival outcomes (patient and graft survival). Independent variables entered to the regression models predicting post-LT outcomes included patient age, MELD score at time of LT, pre-LT RRT, TR, RV FAC, TAPSE, RV/LV ratio, PASP, SBP, RV end-diastolic area, QT length, diastolic parameters (E/A ratio, deceleration time, isovolumetric relaxation time, size of LA, ejection fraction, wall motion thickness and ratio of pre–ejection period to LV ejection time) and donor variables (including age, cold ischemia time and donor type). Candidate covariates were assessed for inclusion into the multivariable model and were selected if p ≤ 0.15 in the univariate results. Throughout the report, confidence interval estimation was using Wald method and statistical significance was established at an alpha level of .05. All statistical analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC).