|
|
Critical Preoperative Data Predict the Need for Concurrent LVAD and RVAD Implantation, Avoiding Poor Outcomes Associated with Delayed Conversion of LVAD to BiVAD
J. Fitzpatrick1, J. R. Frederick1, W. Hiesinger1, R. C. McCormick1, V. M. Hsu1, E. D. Kozin1, M. O'Hara1, E. Howell1, J. E. Cohen1, K. W. Southerland1, J. L. Howard1, E. A. Nacke1, C. M. LaPorte2, M. L. Jessup2, R. J. Morris1, M. A. Acker1, Y. Woo1. 1Hospital of the University of Pennsylvania, Philadelphia, PA, 2University of Pennsylvania, Philadelphia, PA,
BACKGROUND: It is generally accepted that patients who require biventricular mechanical support (BiVAD) have poorer outcomes than those requiring isolated left ventricular support (LVAD). However, it is difficult to predict which patients will thrive with LVAD support alone. We hypothesized that immediate BiVAD placement improves survival compared to delayed conversion of LVAD to BiVAD, and thus aimed to identify preoperative predictors of isolated LVAD support failure.
METHODS: We reviewed patient characteristics, laboratory and hemodynamic data, and outcomes of 266 patients undergoing LVAD ± RVAD placement at the University of Pennsylvania from April 1995 to June 2007. We subdivided BiVAD patients into immediate BiVAD (I-BiVAD) and delayed BiVAD (D-BiVAD) groups. We defined D-BiVAD as failure of isolated LVAD support, even if failure consisted of an attempt to wean cardiopulmonary bypass and subsequent RVAD insertion during the same operation.
RESULTS: Of 266 LVAD patients, 99 required BiVAD (37%). We compared 36 preoperative criteria between LVAD and BiVAD groups to identify risk factors for failure of isolated LVAD support. Univariate analysis showed statistical significance in 23 variables. By multivariate logistic regression, low cardiac index (CI), low right ventricular stroke work index (RVSWI), severe preoperative RV dysfunction, elevated creatinine, prior cardiac surgery, and low systolic blood pressure (SBP) were most significant, and when built into a weighted algorithm were highly predictive of isolated LVAD support failure.
We then compared successful bridging to transplant (excluding recovery and destination patients), survival to hospital discharge, and Kaplan-Meier survival between LVAD, I-BiVAD, and D-BiVAD groups. LVAD outcomes exceeded both I-BiVAD and D-BiVAD outcomes (Figure 1). Further, I-BiVAD (n=71) patients had superior survival to discharge than D-BiVAD (n=28) patients (51% v 29% p<0.05). Kaplan-Meier one-year survival confirmed this finding (48% v 25% p<0.02). There was also a trend towards improved bridging to transplant in I-BiVAD (n=55) vs. D-BiVAD (n=22) patients (65% v 45% p=0.10).
CONCLUSIONS: Failure of LVAD support alone can be reliably predicted using preoperative CI, RVSWI, preoperative RV dysfunction, creatinine, prior cardiac surgery, and SBP. When patients likely requiring BiVAD are identified, proceeding directly to biventricular support is advised, as even brief failure of isolated LVAD support dramatically decreases survival.
Back to 2008 Annual Meeting
Back to Program Outline
|
