Grimm Joshua C, Lui Cecillia, Kilic Arman, Valero Vicente, Sciortino Christopher M, Whitman Glenn J R, Shah Ashish S
Division of Cardiac Surgery, The Johns Hopkins Hospital, Baltimore, Maryland.
Division of Cardiac Surgery, The Johns Hopkins Hospital, Baltimore, Maryland.
Ann Thorac Surg. 2015 Jan;99(1):251-7. doi: 10.1016/j.athoracsur.2014.07.073. Epub 2014 Nov 14.
Despite the significant morbidity associated with renal failure after lung transplantation (LTx), no predictive models currently exist. Accordingly, the purpose of this study was to develop a preoperative risk score based on recipient-, donor-, and transplant-specific characteristics to predict postoperative acute renal failure in candidates for transplantation.
The United Network of Organ Sharing (UNOS) database was queried for adult patients (≥ 18 years of age) undergoing LTx between 2005 and 2012. The population was randomly divided into derivation (80%) and validation (20%) cohorts. The primary outcome of interest was new-onset renal failure. Variables predictive of acute renal failure (exploratory p value < 0.2) within the derivation cohort were incorporated into a multivariable logistic regression model. Odds ratios were used to assign values to the independent predictors of postoperative renal failure to construct the risk stratification score (RSS).
During the study period, 10,963 patients underwent lung transplantation, and the incidence of renal failure was 5.5% (598 patients). Baseline recipient-, donor-, and transplant-related factors were similar between the cohorts. Eighteen covariates were included in the multivariable model, and 10 were assigned values based on their relative odds ratios (ORs). Scores were stratified into 3 groups, with an observed rate of acute renal failure of 3.1%, 5.3%, and 15.6% in the low-, moderate-, and high-risk groups, respectively. The incidence of renal failure was found to be significantly increased in the highest risk group (p < 0.001). Furthermore, the risk model's predicted rates of renal failure highly correlated with actual rates observed in the population (r = 0.86).
We introduce a novel and simple RSS that is highly predictive of renal failure after LTx.
尽管肺移植（LTx）后肾衰竭会导致严重的发病率，但目前尚无预测模型。因此，本研究的目的是基于受者、供者和移植特异性特征开发一种术前风险评分，以预测移植候选者术后急性肾衰竭。
查询器官共享联合网络（UNOS）数据库中2005年至2012年间接受LTx的成年患者（≥18岁）。将该人群随机分为推导队列（80%）和验证队列（20%）。感兴趣的主要结局是新发肾衰竭。将推导队列中预测急性肾衰竭的变量（探索性p值<0.2）纳入多变量逻辑回归模型。使用比值比为术后肾衰竭的独立预测因素赋值，以构建风险分层评分（RSS）。
在研究期间，10963例患者接受了肺移植，肾衰竭发生率为5.5%（598例患者）。各队列之间的基线受者、供者和移植相关因素相似。多变量模型纳入了18个协变量，其中10个根据其相对比值比（OR）赋值。评分分为3组，低、中、高风险组的急性肾衰竭观察发生率分别为3.1%、5.3%和15.6%。发现最高风险组的肾衰竭发生率显著增加（p<0.001）。此外，风险模型预测的肾衰竭发生率与人群中观察到的实际发生率高度相关（r=0.86）。
我们引入了一种新颖且简单的RSS，它能高度预测LTx后的肾衰竭。
