Readmission due to driveline infection can be predicted by new score by using serum albumin and body mass index during long-term left ventricular assist device support.

Survival in patients with continuous flow left ventricular assist device (CF LVAD) had been increased owing to improved perioperative management procedures. The second target for successful long-term LVAD treatment was to reduce readmission especially due to device-specific infection, which was one of the major unsolved complications. Among 57 enrolled patients who had received CF LVAD and been followed for 530 days on median at our institute between 2008 and 2014, 21 patients experienced readmission due to driveline infection (DLI) at 190 days after the surgery on median. Considering the result of Uni/Multivariate Cox regression analyses demonstrating lower serum albumin concentration (S-ALB) (hazard ratio 0.144) and body mass index (BMI) (hazard ratio 0.843) both obtained at discharge were independent predictors of readmission due to DLI, we constructed a New Score "7 × [S-ALB (g/dL)] + [BMI]", which significantly stratified readmission-free rate into 3 groups [low (>50 Pt), intermediate (44-50 Pt), and high risk group (<44 Pt)] during 2-year study period (p = 0.008). Survival remained unchanged irrespective of DLI, whereas those with DLI needed longer in-hospital treatment (p < 0.05). In conclusion, readmission due to DLI could be predicted by using two simple nutrition parameters at discharge. Early nutrition assessment and intervention may reduce readmission and improve patients' quality of life during long-term LVAD support.