Development of a predictive model for long-term survival after lung transplantation and implications for the lung allocation score.
Gries C J, Rue TC, et al.Journal of Heart & Lung Transplantation 2010; 29(7): 731-738.
Aims
To develop and validate a predictive survival model for patients up to five years after transplantation using pre-transplant characteristics.
Interventions
Cases were split into either a development or validation dataset.
Participants
18072 adult lung transplant recipients.
Outcomes
The outcomes included baseline characteristics of patients; age, gender, diagnosis, creatinine, bilirubin, oxygen requirements, cardiac output, Epstein-Barr viral status, transfusion history and diabetes history to inform survival prediction models (model conditioned on surviving to one year, models prediction 5 year survival for separate disease sub-groups, and the predictive availability of the lung allocation system (LAS) post-transplant survival model for 1 and 5 year survival).
Follow-up
N/A
CET Conclusions
This study is not truly a randomised controlled trial; rather a large analysis of registry data from the ISHLT in an attempt to create a model to predict the survival of lung transplant recipients from pre-operative patient characteristics. The cohort of 18,072 cases was randomly split into development and validation datasets (although how this was done is not described). Neither an overall model, nor separate models for disease subgroups (e.g. COPD, idiopathic fibrosis, cystic fibrosis), performed adequately to predict survival in the validation cohort. The usual caveats about database analyses of this type hold: analysis may be affected by missing data or inaccurate data due to the retrospective nature of data entry. It is possible that the results here are simply a demonstration of the importance of donor characteristics on transplant outcomes, and that models based upon recipient data alone will not accurately predict survival.
Quality notes
Quality assessment not appropriate – development and validation.
Trial registration
Not reported.