Faculty Bibliography

BACKGROUND & AIMS:The Model for End-Stage Liver Disease (MELD) has been established as a reliable indicator of short-term survival in patients with end-stage liver disease. The current version (MELDNa), consisting of the international normalized ratio and serum bilirubin, creatinine, and sodium, has been used to determine organ allocation priorities for liver transplantation in the United States. The objective was to optimize MELD further by taking into account additional variables and updating coefficients with contemporary data. METHODS:All candidates registered on the liver transplant wait list in the US national registry from January 2016 through December 2018 were included. Uni- and multivariable Cox models were developed to predict survival up to 90 days after wait list registration. Model fit was tested using the concordance statistic (C-statistic) and reclassification, and the Liver Simulated Allocation Model was used to estimate the impact of replacing MELDNa with the new model. RESULTS:The final multivariable model was characterized by (1) additional variables of female sex and serum albumin, (2) interactions between bilirubin and sodium and between albumin and creatinine, and (3) an upper bound for creatinine at 3.0 mg/dL. The final model (MELD 3.0) had better discrimination than MELDNa (C-statistic, 0.869 vs 0.862; P < .01). Importantly, MELD 3.0 correctly reclassified a net of 8.8% of decedents to a higher MELD tier, affording them a meaningfully higher chance of transplantation, particularly in women. In the Liver Simulated Allocation Model analysis, MELD 3.0 resulted in fewer wait list deaths compared to MELDNa (7788 vs 7850; P = .02). CONCLUSION:MELD 3.0 affords more accurate mortality prediction in general than MELDNa and addresses determinants of wait list outcomes, including the sex disparity.

MELD-Na appears to disadvantage women awaiting liver transplant by underestimating their mortality rate. Fixing this problem involves: (1) estimating the magnitude of this disadvantage separately for each MELD-Na, (2) designing a correction for each MELD-Na, and (3) evaluating corrections to MELD-Na using simulated allocation. Using Kaplan-Meier modeling, we calculated 90-day without-transplant survival for men and women, separately at each MELD-Na. For most scores between 15 and 35, without-transplant survival was higher for men by 0-5 percentage points. We tested two proposed corrections to MELD-Na (MELD-Na-MDRD and MELD-GRAIL-Na), and one correction we developed (MELD-Na-Shift) to target the differences we quantified in survival across the MELD-Na spectrum. In terms of without-transplant survival, MELD-Na-MDRD overcorrected sex differences while MELD-GRAIL-Na and MELD-Na-Shift eliminated them. Estimating the impact of implementing these corrections with the liver simulated allocation model, we found that MELD-Na-Shift alone eliminated sex disparity in transplant rates (p = 0.4044) and mortality rates (p = 0.7070); transplant rates and mortality rates were overcorrected by MELD-Na-MDRD (p = 0.0025, p = 0.0006) and MELD-GRAIL-Na (p = 0.0079, p = 0.0005). We designed a corrected MELD-Na that eliminates sex disparities in without-transplant survival, but allocation changes directing smaller livers to shorter candidates may also be needed to equalize women's access to liver transplant.

Recently, model for end-stage liver disease (MELD)-based liver allocation in the United States has been questioned based on concerns that waitlist mortality for a given biologic MELD (bMELD), calculated using laboratory values alone, might be higher at certain centers in certain locations across the country. Therefore, we aimed to quantify the center-level variation in bMELD-predicted mortality risk. Using Scientific Registry of Transplant Recipients (SRTR) data from January 2015 to December 2019, we modeled mortality risk in 33 260 adult, first-time waitlisted candidates from 120 centers using multilevel Poisson regression, adjusting for sex, and time-varying age and bMELD. We calculated a "MELD correction factor" using each center's random intercept and bMELD coefficient. A MELD correction factor of +1 means that center's candidates have a higher-than-average bMELD-predicted mortality risk equivalent to 1 bMELD point. We found that the "MELD correction factor" median (IQR) was 0.03 (-0.47, 0.52), indicating almost no center-level variation. The number of centers with "MELD correction factors" within ±0.5 points, and between ±0.5-± 1, ±1.0-±1.5, and ±1.5-±2.0 points was 62, 41, 13, and 4, respectively. No centers had waitlisted candidates with a higher-than-average bMELD-predicted mortality risk beyond ±2 bMELD points. Given that bMELD similarly predicts waitlist mortality at centers across the country, our results support continued MELD-based prioritization of waitlisted candidates irrespective of center.

The SRTR maintains the liver-simulated allocation model (LSAM), a tool for estimating the impact of changes to liver allocation policy. Integral to LSAM is a model that predicts the decision to accept or decline a liver for transplant. LSAM implicitly assumes these decisions are made identically for adult and pediatric liver transplant (LT) candidates, which has not been previously validated. We applied LSAM's decision-making models to SRTR offer data from 2013 to 2016 to determine its efficacy for adult (≥18) and pediatric (<18) LT candidates, and pediatric subpopulations-teenagers (≥12 to <18), children (≥2 to <12), and infants (<2)-using the area under the receiver operating characteristic (ROC) curve (AUC). For nonstatus 1A candidates, all pediatric subgroups had higher rates of offer acceptance than adults. For non-1A candidates, LSAM's model performed substantially worse for pediatric candidates than adults (AUC 0.815 vs. 0.922); model performance decreased with age (AUC 0.898, 0.806, 0.783 for teenagers, children, and infants, respectively). For status 1A candidates, LSAM also performed worse for pediatric than adult candidates (AUC 0.711 vs. 0.779), especially for infants (AUC 0.618). To ensure pediatric candidates are not unpredictably or negatively impacted by allocation policy changes, we must explicitly account for pediatric-specific decision making in LSAM.

BACKGROUND AND AIMS:Scores from the Model for End-Stage Liver Disease (MELD), which are used to prioritize candidates for deceased donor livers, are widely acknowledged to be negatively correlated with the 90-day survival rate without a liver transplant. However, inconsistent and outdated estimates of survival probabilities by MELD preclude useful applications of the MELD score. APPROACH AND RESULTS:Using data from all prevalent liver waitlist candidates from 2016 to 2019, we estimated 3-day, 7-day, 14-day, 30-day, and 90-day without-transplant survival probabilities (with confidence intervals) for each MELD score and status 1A. We used an adjusted Kaplan-Meier model to avoid unrealistic assumptions and multiple observations per person instead of just the observation at listing. We found that 90-day without-transplant survival has improved over the last decade, with survival rates increasing >10% (in absolute terms) for some MELD scores. We demonstrated that MELD correctly prioritizes candidates in terms of without-transplant survival probability but that status 1A candidates' short-term without-transplant survival is higher than that of MELD 40 candidates and lower than that of MELD 39 candidates. Our primary result is the updated survival functions themselves. CONCLUSIONS:We calculated without-transplant survival probabilities for each MELD score (and status 1A). The survival function is an invaluable tool for many applications in liver transplantation: awarding of exception points, calculating the relative demand for deceased donor livers in different geographic areas, calibrating the pediatric end-stage liver disease score, and deciding whether to accept an offered liver.

BACKGROUND AND AIMS:In February 2020, the Organ Procurement and Transplantation Network replaced donor service area-based allocation of livers with acuity circles, a system based on three homogeneous circles around each donor hospital. This system has been criticized for neglecting to consider varying population density and proximity to coast and national borders. APPROACH AND RESULTS:Using Scientific Registry of Transplant Recipients data from July 2013 to June 2017, we designed heterogeneous circles to reduce both circle size and variation in liver supply/demand ratios across transplant centers. We weighted liver demand by Model for End-Stage Liver Disease (MELD)/Pediatric End-Stage Liver Disease (PELD) because higher MELD/PELD candidates are more likely to be transplanted. Transplant centers in the West had the largest circles; transplant centers in the Midwest and South had the smallest circles. Supply/demand ratios ranged from 0.471 to 0.655 livers per MELD-weighted incident candidate. Our heterogeneous circles had lower variation in supply/demand ratios than homogeneous circles of any radius between 150 and 1,000 nautical miles (nm). Homogeneous circles of 500 nm, the largest circle used in the acuity circles allocation system, had a variance in supply/demand ratios 16 times higher than our heterogeneous circles (0.0156 vs. 0.0009) and a range of supply/demand ratios 2.3 times higher than our heterogeneous circles (0.421 vs. 0.184). Our heterogeneous circles had a median (interquartile range) radius of only 326 (275-470) nm but reduced disparities in supply/demand ratios significantly by accounting for population density, national borders, and geographic variation of supply and demand. CONCLUSIONS:Large homogeneous circles create logistical burdens on transplant centers that do not need them, whereas small homogeneous circles increase geographic disparity. Using carefully designed heterogeneous circles can reduce geographic disparity in liver supply/demand ratios compared with homogeneous circles of radius ranging from 150 to 1,000 nm.

Recently, the Organ Procurement and Transplant Network approved a plan to allocate kidneys within 250-nm circles around donor hospitals. These homogeneous circles might not substantially reduce geographic differences in transplant rates because deceased donor kidney supply and demand differ across the country. Using Scientific Registry of Transplant Recipients data from 2016-2019, we used an integer program to design unique, heterogeneous circles with sizes between 100 and 500 nm that reduced supply/demand ratio variation across transplant centers. We weighted demand according to wait time because candidates who have waited longer have higher priority. We compared supply/demand ratios and average travel distance of kidneys, using heterogeneous circles and 250 and 500-nm fixed-distance homogeneous circles. We found that 40% of circles could be 250 nm or smaller, while reducing supply/demand ratio variation more than homogeneous circles. Supply/demand ratios across centers for heterogeneous circles ranged from 0.06 to 0.13 kidneys per wait-year, compared to 0.04 to 0.47 and 0.05 to 0.15 kidneys per wait-year for 250-nm and 500-nm homogeneous circles, respectively. The average travel distance for kidneys using heterogeneous, and 250-nm and 500-nm fixed-distance circles was 173 nm, 134 nm, and 269 nm, respectively. Heterogeneous circles reduce geographic disparity compared to homogeneous circles, while maintaining reasonable travel distances.

BACKGROUND:In December 2018, United Network for Organ Sharing approved an allocation scheme based on recipients' geographic distance from a deceased donor (acuity circles [ACs]). Previous analyses suggested that ACs would reduce waitlist mortality overall, but their impact on pediatric subgroups was not considered. METHODS:We applied Scientific Registry of Transplant Recipients data from 2011 to 2016 toward the Liver Simulated Allocation Model to compare outcomes by age and illness severity for the United Network for Organ Sharing-approved AC and the existing donor service area-/region-based allocation schemes. Means from each allocation scheme were compared using matched-pairs t tests. RESULTS:During a 3-year period, AC allocation is projected to decrease waitlist deaths in infants (39 versus 55; P < 0.001), children (32 versus 50; P < 0.001), and teenagers (15 versus 25; P < 0.001). AC allocation would increase the number of transplants in infants (707 versus 560; P < 0.001), children (677 versus 547; P < 0.001), and teenagers (404 versus 248; P < 0.001). AC allocation led to decreased median pediatric end-stage liver disease/model for end-stage liver disease at transplant for infants (29 versus 30; P = 0.01), children (26 versus 29; P < 0.001), and teenagers (26 versus 31; P < 0.001). Additionally, AC allocation would lead to fewer transplants in status 1B in children (97 versus 103; P = 0.006) but not infants or teenagers. With AC allocation, 77% of pediatric donor organs would be allocated to pediatric candidates, compared to only 46% in donor service area-/region-based allocation (P < 0.001). CONCLUSIONS:AC allocation will likely address disparities for pediatric liver transplant candidates and recipients by increasing transplants and decreasing waitlist mortality. It is more consistent with federally mandated requirements for organ allocation.