The UNOS Standard Transplant Analysis and Research (STAR) database represents an open cohort of prospectively collected data involving all United States patients receiving LTx from 1987 until December 2008, with follow-up extending until September 2009. Our institution deemed IRB approval unnecessary as no patient or center identifiers were included in this analysis.
This study was a retrospective cohort design, including adult (>17 years) patients undergoing LTx in the post-LAS era (March 2005-present). Exclusion criteria included incomplete preservation solution information, heart-lung transplantation, and patients with prior LTx. The cohort was stratified according to whether donor lungs were preserved in LPD (LDS) or University of Wisconsin (UW) solution.
Variables Examined and Outcome Measures
Pertinent variables examined within the dataset included: demographic factors (age, gender, race, and education level); markers of pulmonary status (oxygen requirement, six minute walking distance, forced expiratory volume at 1 second (FEV1), forced vital capacity (FVC), FEV1 to FVC ratio, mechanical ventilation prior to LTx, and ICU care prior to transplant); co-morbidities (LAS score, diabetes mellitus, body mass index (BMI), preoperative creatinine levels and hypertension); and transplant variables (ischemic time, HLA mismatch, panel reactive antibody (PRA) level, year of transplant, and wait list times). We further examined donor variables including donor age, race, gender, cigarette use, and BMI.
The primary endpoint was the incidence of 1-year mortality. Secondary outcomes examined were short term mortality (30-day and 90-day), as well as rejection requiring treatment within the first year following LTx.
We compared baseline characteristics among the LPD and UW groups by the student’s t-test (continuous variables) and the chi-square test (categorical variables). 30-day, 90-day, and 1-year survival were estimated using the Kaplan-Meier method, as these time intervals have adequate follow-up in the post-LAS era. To compare survival estimates according to preservation solution, the Mantel-Cox log-rank test was used. The entire cohort was analyzed according to the Kaplan-Meier method. Separate Kaplan-Meier analysis was performed in the upper two quartiles of LAS to assess the impact of preservation solution in high risk patients.
A multivariable Cox proportional hazards regression model estimated risk of death with censoring for death, loss to follow-up, and administrative reasons. To construct the multivariable model, independent covariates with potential for confounding were first tested in a univariate fashion. In addition to variables associated with mortality on exploratory analysis (p<0.1), those with biological plausibility and previously recognized risk factors were incorporated in a forwards and backwards stepwise fashion into the multivariable model. The likelihood ratio test and Akaike’s information criterion in a nested model approach were used to identify which covariates increased the explanatory power of the model. As the multivariable model was developed with case-wise deletion, all covariates with greater than 15% missing data were not included. The final model incorporated the following covariates: storage solution, recipient age ≥ 65, creatinine level, ICU prior to transplant, hospitalization prior to transplant, final LAS calculation, organ ischemic time, donor cigarette use, donor age, and donor CMV status.
For all analyses, a p-value of less than 0.05 (two-tailed) was considered significant. Means are displayed with standard deviations. Hazard ratios are presented with 95% confidence intervals (CI). Statistical testing was performed using STATA software (version 9.2 SE, StataCorp LP, College Station, TX).