Patient selection for lung transplantation.

Abstract

Listen

Translate article

ISOLATED LUNG TRANSPLANTATION WAS FIRST PERformed successfully as a unilateral graft in a patient with idiopathic pulmonary fibrosis in 1983. Beginning around 1990, as the number of transplant centers increased the number of lung transplants also increased rapidly. However, by 1996 the limit of the donor pool was reached and the number of transplants plateaued. Since then, the number of lung transplant operations performed throughout the world has averaged between 1300 and 1400 per year, with approximately 65% performed in the United States. Meanwhile, the United Network for Organ Sharing (UNOS) list of patients with various types of end-stage lung diseases who are awaiting transplantation has steadily grown, and now includes nearly 4000 names. This disparity means that most patients waiting for a lung graft will never receive one. At the same time, favorable outcomes following lung transplantation have lagged behind those of other solid organs. The Registry for the International Society for Heart and Lung Transplantation (ISHLT), which is jointly sponsored by UNOS, reports that patients receiving a lung transplant can expect to have survival rates of approximately 75% at 1 year and 45% at 5 years. Over the past 12 years, these rates have improved only slightly. With the limited availability of lung grafts and suboptimal recipient survival, the process of selection of surgical candidates becomes a critical factor. Appropriate candidates should be sufficiently ill to benefit from transplantation, yet well enough to survive the rigors of the procedure. Selecting appropriate candidates for lung transplantation is difficult. Longitudinal epidemiologic studies involving large groups of patients with progressive, fatal pulmonary processes are necessary to identify prognostic factors, but such studies are difficult to perform and are few in number. In addition, an individual’s unique medical characteristics often influence disease progression, thereby making it difficult to apply statistical models to specific patients. When International Guidelines for the Selection of Candidates for Lung Transplantation were published in 1998, one major study, by Kerem et al, had been published that identified prognostic factors in patients with cystic fibrosis (CF). This study identified FEV1 (and, to a lesser extent, several other parameters) as a useful marker for predicting survival; an FEV1 of less than 30% of the predicted value was associated with 2-year survival of approximately 50%. However, the study was limited by only including patients from 1 clinical center and, therefore, might have been flawed by biases related to the patient population or treatment practices of that particular center. For example, the patient population had a high rate of colonization with Burkholderia cepacia, an organism known to influence survival in patients with advanced pulmonary disease. Nevertheless, the International Guidelines Committee adopted criteria from the study by Kerem et al as the best available at that time for selection of patients with CF who are candidates for lung transplantation. Several subsequent studies challenged the use of FEV1 as the major factor for predicting outcomes following lung transplantation for CF, although most included relatively small patient populations. Some studies used multivariable regression analysis to define significant variables associated with mortality and then applied these to develop models to predict survival more precisely in these patients. The model created by Liou et al is particularly appealing because it was developed using the largest collection of CF patient data available, the Cystic Fibrosis Foundation Patient Registry (CFFPR). The authors used data from more than 5800 registry patients to develop the 9-parameter model and then validated the model in a second similarly sized group of patients. After applying their model, the authors stratified patients into 5 categories based on predicted survival ranging from less than 30% to approaching 100%, and then asked whether the model would perform better than criteria used previously for choosing appropriate candidates for transplantation. In this issue of THE JOURNAL, Liou and colleagues now attempt to answer that question. The authors obtained pretransplant clinical information from the CFFPR for 468 patients who had subsequently received lung transplants from 1992 through 1997 and compared their posttransplant survival over 5 years with predicted survival for 11630 pa-