Background: Cardiac complications are the leading cause of long-term non-graft related mortality following liver transplantation. While the Scientific Registry of Transplant Recipients (SRTR) reports multiple risk factors for post-operative mortality, there are currently no cardiac specific risk factors included in the survival models. Cirrhotic cardiomyopathy, a spectrum of cardiovascular changes associated with end stage liver disease including diastolic dysfunction (DD), is well described. Left atrial volume index (LAVI) is an echocardiographic measure of DD, and has been associated with mortality in many populations. Unlike transmitral inflow velocities which are load dependent, LAVI is considered to be a more sensitive and stable measurement of DD. To date, there is limited data evaluating the effect of LAVI on mortality following liver transplantation. The aims of this study were to determine whether LAVI is an independent predictor of post-liver transplant mortality, and whether LAVI improves the ability to predict mortality beyond known risk factors.
Methods: We performed a retrospective cohort study of patients ≥ 18 years of age who underwent liver transplantation between July 2011 and June 2014 at the University of California, Los Angeles, and who had their preoperative transthoracic echocardiograms performed at our center. The primary outcome was time to mortality, and the primary predictor was LAVI, dichotomized at 28ml/m2. Known risk factors of post liver transplant mortality identified from the SRTR database were collected as covariates. A multivariable Cox regression model was built using a backwards stepwise selection procedure to assess the effect of LAVI on post-operative mortality.
Results: Of the 254 patients included in our analysis, 48 deaths occurred over the follow-up period (median: 17.5 months). In a multivariable model including re-transplantation, physiologic MELD score (dichotomized at the sample median of 33), preoperative mechanical ventilation, previous malignancy, and HCV, LAVI was not a statistically significant predictor of mortality (HR: 0.99, p=0.99, 95% CI: 0.56, 1.77). Given that advanced liver disease is associated with cirrhotic cardiomyopathy, we explored whether the effect of LAVI on mortality differed as a function of MELD score. In a multivariable model including the covariates listed above, there was a statistically significant interaction between LAVI and MELD score (p=0.007). Specifically, for patients with MELD scores ≥ 33, LAVI ≥ 28 ml/m2 was associated with increased mortality (HR=2.4, p=0.032, 95% CI 1.1, 5.4). However, for patients with MELD scores < 33, LAVI was not associated with mortality (HR: 0.44, p=0.08, 95% CI 0.18, 1.1). The C-statistic for the model including LAVI and the interaction was 0.73, statistically significantly greater than the C-statistic of 0.67 for the model excluding these terms, thus demonstrating an improvement in the predictive ability of that model.
Discussion: This is the first study to examine the effect of LAVI as a predictor of post-liver transplant mortality using a multivariable model. We demonstrated that LAVI had a significant impact on mortality among patients with high MELD scores, whereas this effect was not observed among patients with lower MELD scores. Liver transplant recipients with high LAVI values and high MELD scores may represent patients with advanced cirrhotic cardiomyopathy who may be at an increased risk of postoperative mortality. This may have important consequences for the selection of liver transplant recipients.