A new medical urgency score may better identify adult heart transplant candidates at higher risk of mortality without transplantation than the current US 6-status allocation system.
Investigators analyzed data from the Scientific Registry of Transplant Recipients involving 33,097 adult heart transplant candidates on the waiting list between January 2019 and August 2025. The population was divided into training (n. 21,859), validation (n = 7,897), and test (n = 8,256) data sets, with some patients contributing data to more than 1 set.
The investigators assessedmortality within 6 weeks without transplantation, defined as mortality before transplantation or removal from the waiting list because of clinical deterioration. They developed the US Candidate Risk Score (US-CRS) 2.0 using a marginal structural model designed to account for differences between candidates who underwent transplantation and those remaining on the waiting list. The score incorporated 5 laboratory measurementsand 4 categories of mechanical circulatory support. Model performance was assessed through discrimination and calibration analyses, followed by simulations comparing allocation priorities under the proposed and existing systems.
In the test data set, the simpler US-CRS 2.0 model achieved a mean area under the receiver operating characteristic curve of 0.807 for predicting 6-week mortality without transplantation vs. 0.693 for the existing 6-status system. The more complex model achieved a mean value of 0.795. The simpler model demonstrated better discrimination than the existing system with waiting-time tiebreakers in 3 of 8 evaluation periods.
Both versions of the score were well calibrated across urgency statuses, exception categories, and diagnoses but overestimated mortality among candidates at the highest predicted risk, with greater overestimation in the complex model. Predictive performance remained similar after converting the model into a 50-point allocation score, using multiple imputation for missing training data, and excluding patients whose records appeared in training and test sets.
In an analysis of 45,530 historical donor heart match runs from 2019 through 2024, the investigators evaluated 2 proposed allocation approaches:replacing existing adult urgency statuses with categories derived from US-CRS while retaining other allocation considerations and ranking adult candidates nationally according to their 50-point scores.
Under the proposed 6-status US-CRS approach, observed 6-week mortality among candidates ranked first for a donor heart was 16% vs. 11% under the existing system. With the national US-CRS approach, mortality among first-ranked candidates reached 23%, indicating that both proposed approaches placed candidates with greater short-term mortality risk at the top of the allocation list.
The revised allocation approaches changed candidate rankings. Under the 6-status US-CRS system, 57% of transplant recipients moved at least 10 positions in the match-run sequence, and 67% moved to lower-priority positions. Under the national US-CRS approach, 80% moved at least 10 positions. The 6-status US-CRS approach reduced donor-to-candidate distances at earlier match-run positions, whereas the national approach increased those distances.
Additional analyses identified variation in predicted mortality among candidates assigned the same urgency status under the existing system. Among status 4 candidates, those not using durable left ventricular assist devices had a median urgency score of 30 vs. 24 among those using these devices. The investigators found variation among candidates receiving exception statuses, indicating differences in predicted mortality not fully reflected by existing categories.
The score differentiated mortality risk according to mechanical circulatory support. Durable left ventricular assist device use was associated with lower short-term mortality, whereas intra-aortic balloon pumps, microaxial flow pumps, and advanced mechanical circulatory support were associated with higher mortality. Advanced mechanical circulatory support demonstrated the strongest association with mortality among the device categories evaluated.
The investigators evaluated the new score's susceptibility to manipulation through device selection. For most candidates not receiving mechanical circulatory support, initiating an intra-aortic balloon pump or microaxial flow pump resulted in either no change or an increase of 1 urgency category. By comparison, initiation of either device could move a candidate from status 6 to status 2 under the existing allocation system. The findings suggested that device initiation alone would generally be insufficient to achieve a high-priority US-CRS classification.
Subgroup analyses identified changes in allocation priority across candidate groups. Both approaches prioritized more candidates from racial and ethnic minority groups at earlier match-run positions while reducing priority for candidates with blood type O and those using durable left ventricular assist devices. Changes in priority by sex and stature were minimal.
The study had several limitations. The model overestimated mortality among candidates at the highest predicted risk and would require periodic recalibration as transplant practices and heart failure management evolve. The match-run analysis could not quantify potential effects on survival after transplantation if donor hearts were allocated to candidates with greater medical urgency. The reduced priority assigned to candidates with blood type O and those using durable left ventricular assist devices highlighted the need for separate allocation provisions to maintain equitable access. The investigators emphasized that a full simulation incorporating these considerations would be necessary before implementation.
“Replacing the current 6 urgency statuses with the more accurate US-CRS could lead to heart allocation that is less susceptible to manipulation and saves more lives,” wrote lead study author Grace R. Lyden, PhD, of the Scientific Registry of Transplant Recipients at Hennepin Healthcare Research Institute and the University of Minnesota, and colleagues.
Full disclosures of the study authors can be found in the study.
Source: JAMA
