David L Joyce1, Zhuo Li2, Leah B Edwards3, Jon A Kobashigawa4, Richard C Daly5. 1. Department of Cardiovascular Surgery, Mayo Clinic, Rochester, Minn. 2. Biomedical Statistics and Informatics, Mayo Clinic, Jacksonville, Fla. 3. United Network for Organ Sharing, Richmond, Va. 4. Cedars-Sinai Heart Institute, Los Angeles, Calif. 5. Department of Cardiovascular Surgery, Mayo Clinic, Rochester, Minn. Electronic address: rdaly@mayo.edu.
Abstract
OBJECTIVE: Many donor and recipient factors influence 1-year survival of patients after cardiac transplantation. To date, a statistical model has not been developed to assess the interplay of these factors in predicting outcomes, so we developed a risk-assessment tool to enhance decision-making. METHODS: We analyzed 29 variables that were reported in the United Network for Organ Sharing database for 24,540 cardiac transplantations performed between January 1, 2000, and June 30, 2015. For one half of the patients (the prediction population), a multivariable Cox regression model and the bootstrap resampling method were used to devise a scoring system predicting 1-year survival. The other half (the validation population) were stratified by score into 3 risk categories: high risk, medium risk, and low risk. One-year survival was compared among the 3 groups. RESULTS: Eleven variables were statistically significant in predicting 1-year survival. One-year survival for patients with risk scores of less than or equal to 8, 9 to 15, and greater than 15 were 91.2%, 81.7%, and 64.6%, respectively (P < .001). The C index of the Cox regression model was calculated at 0.62 when using risk score as a continuous predictor. CONCLUSIONS: Donor and recipient risk factors influence patient survival after cardiac transplantation. Long-term outcomes may be optimized with a statistically based risk model to improve donor-recipient matching.
OBJECTIVE: Many donor and recipient factors influence 1-year survival of patients after cardiac transplantation. To date, a statistical model has not been developed to assess the interplay of these factors in predicting outcomes, so we developed a risk-assessment tool to enhance decision-making. METHODS: We analyzed 29 variables that were reported in the United Network for Organ Sharing database for 24,540 cardiac transplantations performed between January 1, 2000, and June 30, 2015. For one half of the patients (the prediction population), a multivariable Cox regression model and the bootstrap resampling method were used to devise a scoring system predicting 1-year survival. The other half (the validation population) were stratified by score into 3 risk categories: high risk, medium risk, and low risk. One-year survival was compared among the 3 groups. RESULTS: Eleven variables were statistically significant in predicting 1-year survival. One-year survival for patients with risk scores of less than or equal to 8, 9 to 15, and greater than 15 were 91.2%, 81.7%, and 64.6%, respectively (P < .001). The C index of the Cox regression model was calculated at 0.62 when using risk score as a continuous predictor. CONCLUSIONS:Donor and recipient risk factors influence patient survival after cardiac transplantation. Long-term outcomes may be optimized with a statistically based risk model to improve donor-recipient matching.
Authors: Michiel Morshuis; Sebastian V Rojas; Kavous Hakim-Meibodi; Artyom Razumov; Jan F Gummert; René Schramm Journal: Ann Cardiothorac Surg Date: 2020-03
Authors: Rene Schramm; Armin Zittermann; Uwe Fuchs; Jan Fleischhauer; Angelika Costard-Jäckle; Maria Ruiz-Cano; Luminata-Adriana Krenz; Henrik Fox; Julia Götte; Sabina P W Günther; Stefan Wlost; Sebastian V Rojas; Kavous Hakim-Meibodi; Michiel Morshuis; Jan F Gummert Journal: ESC Heart Fail Date: 2021-10-26