Liver allocation policy recently was modified to use the Model for End-Stage Liver Disease (MELD) for patients with chronic liver disease to stratify potential recipients according to risk for waitlist death. In this study, a retrospective cohort of 760 adult patients with chronic liver disease placed on the liver transplant waitlist between January 1995 and March 2001 and followed up for up to 74 months was studied to assess the ability of the MELD to predict mortality among waitlisted candidates and evaluate the prognostic importance of changes in MELD score over time. Serial MELD scores predicted waitlist mortality significantly better than baseline MELD scores or medical urgency status. Each unit of the 40-point MELD score was associated with a 22% increased risk for waitlist death (P <.001), whereas medical urgency status was not a significant independent predictor. For any given MELD score, the magnitude and direction of change in MELD score during the previous 30 days (DeltaMELD) was a significant independent mortality predictor. Patients with MELD score increases greater than 5 points over 30 days had a threefold greater waitlist mortality risk than those for whom MELD scores increased more gradually (P <.0001). We conclude that mortality risk on the liver transplant waitlist is predicted more accurately by serial MELD score determinations than by medical urgency status or single MELD measurements. DeltaMELD score over time reflects progression of liver disease and conveys important additional prognostic information that should be considered in the further evolution of national liver allocation policy.