Background: As we have previously reported, the preoperative profile defined by INTERMACS is a good predictor for the prognosis after left ventricular assist device (LVAD) implantation, but is largely dependent on the physician's decision. Several other risk stratification systems including objective parameters (eg, Leitz-Miller, Columbia, Seattle Heart Failure Model, APACHE II) have been proposed to estimate patient's mortality after LVAD implantation.
Methods and results: According to the preoperative data from 59 patients who received LVAD (10 implantable, 49 extracorporeal) since 2002 through 2010, we performed a logistic analysis and constructed a new scoring system (ie, the TODAI VAD score (TVAD score), assigning 8 points to serum albumin <3.2mg/dl (odds ratio [OR] 8.475), 7 points to serum total bilirubin >4.8mg/dl (OR 7.300), 6 points to left ventricular end-diastolic diameter <55mm (OR 5.917), 5 points to central venous pressure >11mmHg (OR 5.128)). The receiver-operating characteristic analysis showed that the area under the curve of our new scoring system (0.864) was significantly larger than any of the abovementioned 5 scoring methods (all P<0.05). With the TVAD score, low (0-8 points), intermediate (9-17 points), and high (18-26 points) risk strata had significantly different 1-year survival rates of 95%, 54%, and 14%, respectively (all P<0.001).
Conclusions: The TVAD score can predict the prognosis after LVAD implantation much better than the previously known methods.