Abstract
RATIONALE: Clinical features that impact outcomes in lung recipients with chronic lung allograft dysfunction (CLAD) are uncertain and limited by existing approaches to phenotyping that rely on measures inconsistently performed or not standardized across centers. OBJECTIVE: We used data from the prospective multicenter Lung Transplant Outcomes Group (LTOG) cohort study to determine if routine, objective clinical measures at CLAD diagnosis impact progression to graft loss (death/retransplantation). METHODS: The analysis included 2386 adult lung recipients from 9 US centers. Probable CLAD was defined according to International Society for Heart and Lung Transplantation criteria. Cox models were fit for time from probable CLAD to graft loss as a function of each clinical feature of interest including CLAD stage (1-4), CLAD timing (early-onset, <2 years posttransplant), FVC loss (FVCCLAD/FVCBest <0.8), and FEV1/FVC ratio <0.7. Models included a random effect for center and were adjusted for recipient age, sex, transplant type, and native lung disease. MEASUREMENTS AND MAIN RESULTS: Probable CLAD developed in 1418 patients (59.4%). At onset, 80.7% had stage 1 CLAD, 41.3% had early-onset CLAD, 35.8% had FVC loss, and 46.5% had an FEV1/FVC ratio <0.7. In adjusted analyses, patients with CLAD stage ≥2, early-onset CLAD, and FVC loss had significantly higher hazards for graft loss, while FEV1/FVC <0.7 was associated with reduced graft loss risk. CONCLUSIONS: We provide compelling evidence that routine, objective measures at CLAD onset inform CLAD progression risk. Such stratification of disease behavior is important to patient prognostication and may inform the design of future CLAD trials.