Michelle Kittleson, MD, PhD: Need for Justice in Heart Transplant Allocation

Published on: 

Michelle Kittleson, MD, PhD, provides insight into the potential impact of a new risk score for heart transplant allocation as well as the need for more equitable cardiovascular care in the US.

A new study published in the Journal of the American Medical Association highlights the potential of a novel risk score for determining optimal allocation of donor hearts among transplant candidates in the US.

Developed through an analysis of 46 variables, the new score, named the US candidate risk score or US-CRS, outperformed both the conventional 6-status model currently in use in the US as well as the French-CRS model it was based on.

“I’m delighted, because I think the need is there for a better system that prioritizes the patients who need our help and takes away a physician's imperfect prediction of how to do that for their patients. [Second], the score clearly works. So, if we're moving towards continuous allocation to optimize beneficence and justice, this is a great step towards doing it,” explained Michelle Kittleson, MD, PhD, director of Education in Heart Failure and Transplantation at the Smidt Heart Institute at Cedars-Sinai, in an interview with HCPLive.

The model was built by investigators at the University of Chicago by adding a predefined set of predictors to the current French Candidate Risk Score (French-CRS) model. The final iteration of the US-CRS included time-varying short-term mechanical circulatory support, the log of bilirubin, estimated glomerular filtration rate, the log of B-type natriuretic peptide, albumin, sodium, and durable left ventricular assist device.

Tested in a registry-based observational study of adult heart transplant candidates in the US, assessments of the new score indicated it achieved an AUC for death within 6 weeks of listing of 0.79 (95% Confidence Interval [CI], 0.75-0.83). In comparison, the French-CRS model achieved an AUC of 0.72 (95% CI, 0.67-0.76) and 6-status model achieved an AUC of 0.68 (95% CI, 0.62-0.73). Additional analysis suggested the overall survival concordance were 0.76 (95% CI, 0.73-0.80), 0.69 (95% CI, 0.65-0.73), and 0.67 (95% CI, 0.63-0.71) for the US-CRS model, the French-CRS model, and the 6-status model, respectively.

In an accompanying editorial, Kittleson remarked on the need for greater beneficence and justice in the allocation of donor hearts for transplant candidates. With an interest in obtaining more perspective on the potential of this score to contribute to reform and optimization of donor heart allocation in the US, the editorial team of HCPLive Cardiology reached out to Kittleson for an interview.

Check out the preview video and read the entire Q&A with Kittleson below.

HCPLive: Can you discuss the limitations of the current system and how this new score might contribute to more optimal allocation?

Kittleson: Demand exceeds supply when it comes to donor hearts, so we need a way to allocate them fairly, ensuring that the people who are the sickest receive them first. That is, those who are on the greatest amount of borrowed time, or are closest to experiencing a catastrophic event, will receive a heart first. Thus, the system must be equitable and just, allocating all organs to those who need them the most. There have been iterations of the allocation system since its inception in 1988. The hope with the implementation of the new 6-tier system in October 2018 was that it would address some of the issues of the earlier system, particularly by providing clearer definitions of what constitutes cardiogenic shock. This clarity would ensure that when patients are categorized into the higher urgency statuses (statuses 1 through 3), which are inpatient statuses, it would be evident that it's not merely the physician asserting a patient's condition, but rather demonstrating through specific criteria that the patient is in shock and in need of immediate attention.

So, that was the hope. However, what actually occurred is that there remains room for physician interpretation of the criteria. If a patient meets the criteria—systolic less than 90, index less than two, wedge above 15—they qualify for cardiogenic shock. As the physician caring for the patient in front of you, you still have two choices: temporary mechanical support status or a higher urgency status. There are discrepancies in wait times and levels of urgency between these options. One potential drawback of the current 6-tier system is its susceptibility to physician interpretation. Many dislike the term "manipulation" because it implies that physicians are acting for their own benefit. However, every physician is ultimately advocating for the patient in front of them and considering the welfare of all patients awaiting treatment. It is human nature to prioritize the patient with whom you have a personal relationship. This perceived pitfall has indeed materialized in practice, as evidenced by the increased use of temporary mechanical support devices and exceptions granted to list patients at higher criteria since the system was altered in October 2018. This shift did not signify a sudden increase in the severity of shock requiring balloon pumps and Impellas; rather, it reflected physicians' efforts to advocate for their patients within the available framework.

So, that's why UNOS is endeavoring to transition towards a continuous allocation score, aiming for a more concrete system that minimizes the potential for physician interpretation of criteria. This shift would enable us all to advocate for the patient in front of us without experiencing tension or conflicts regarding the best course of action. A continuous allocation score will incorporate various factors, including medical urgency (the primary focus of the current score), as well as expected post-transplant outcomes, candidate biology, patient access, and efficient management of organ placements. While there are many factors involved, the primary emphasis will remain on medical urgency, with the hope that it will not be susceptible to physician interpretation.

This beautiful study in JAMA addressed the need for a medical urgency score by proposing its own methodology. The researchers analyzed 46 variables from the scientific registry for transplant recipients database, encompassing demographic, clinical, and laboratory parameters, to determine which factors are most predictive of patient outcomes. Their statistical analysis yielded straightforward results: factors such as being on ECMO, having a surgically implanted temporary MCs device, bilirubin levels, EGFR as a measure of kidney function, BNP, albumin levels, sodium levels, and the presence of a durable LVAD were found to be significant predictors. This approach offers several commendable aspects, highlighting its efficacy.

Firstly, it outperformed the current 6-tier allocation system in predicting death on the waiting list. This indicates that although we may believe we are proficient in assigning urgency, there is room for improvement—as demonstrated by this score. Secondly, it is less susceptible to physician interpretation. While it includes variables such as ECMO and surgically implanted MCS, which may seem invasive and carry associated risks, their inclusion reflects the gravity of the patient's condition rather than subjective judgment. Unlike less invasive interventions like balloon pumps, the placement of surgically implanted temporary devices requires significant mental energy and is thus less vulnerable to bias. It is noteworthy that the variable related to balloon pumps did not significantly impact waitlist mortality, highlighting its relatively minor intervention status.

I am pleased because there is a clear need for a more effective system that prioritizes patients in need and mitigates the impact of a physician's imperfect predictions. Additionally, the score's demonstrated effectiveness suggests that it is a viable option as we progress towards continuous allocation to enhance beneficence and justice in organ allocation.

HCPLive: Given the disparities in access to heart transplant among underserved populations, how do efforts like this contribute to more equitable access?

Kittleson: I think the problem starts so much earlier, even back before the patient has reached an evaluation for advanced heart failure therapies. We're truly looking at the tip of the iceberg and we have to focus on what lies beneath. We have to focus on the fact that Black Americans have a higher risk of hypertension, which places them at increased risk for heart failure down the line. They're not getting the same access to the health care so even starting with the primary prevention, the stage a heart failure, already we see biology and disparity combining to lead to a disadvantage.

So, while we may attempt to address the problem at its latest stage, it's evident that access to advanced heart failure care varies. Moreover, studies reveal that even upon gaining access to such care, these patients are less likely to undergo transplantation and more likely to face differing thresholds regarding factors such as social support, caregivers, and adherence. Therefore, I believe we are far from effectively managing this issue. It's imperative that we address the problem as outlined, not only at the advanced stage but also much earlier in the process.

HCPLive: Investigators call specific attention to the absence of hemodynamics from this score, does that surprise you given the recent emphasis on hemodynamic-guided care?

Kittleson: I believe it is implied that optimizing a patient's hemodynamics will lead to improvements in various parameters such as their ability, EGFR, and BNP levels. Consequently, they are less likely to require surgically implantable temporary mechanical devices—or more likely to recover without them. The reason these factors did not explicitly show up in the score may be due to the variability in how they are measured and recorded. While parameters like bilirubin are relatively consistent in measurement, factors like wedge measurements can vary significantly. So, I am actually quite relieved that these measures did not feature prominently in the score. However, the clinically relevant consequences of optimizing perfusion, which is ultimately our goal for patients, did indeed show up in the measure.

HCPLive: Are there any specific limitations to this new score you would want to call out to our audience?

Kittleson: The issue of antibody sensitization remains unresolved in my opinion. Individuals with elevated anti-HLA antibodies, whether due to prior blood transfusions, pregnancies, transplants, or mechanical circulatory support devices, experience longer wait times and increased waitlist mortality. Determining where this factor should fit into a score like this is challenging. Similar to the measurement of the wedge, there are numerous ways to define and encapsulate sensitization, making it difficult to incorporate into a continuous allocation score.

Understanding how sensitization should be integrated into the score will be crucial, especially for historically disadvantaged groups. Perhaps it could be included as a factor in patient access or predicted outcomes. Despite advances in medicine, there will always be a need for exceptions. Every patient is unique, and medical decisions cannot always adhere strictly to guidelines. Artificial intelligence may assist in data analysis, but it cannot replace the need for individualized patient care. The authors may have addressed many aspects of continuous allocation effectively, but the integration of exceptions and considerations for complex issues like sensitization remain open questions in my mind.

Dr. Kittleson has no disclosures of note to report.

Editor’s note: This transcript was edited for grammar and clarity using artificial intelligence tools.


  1. Zhang KC, Narang N, Jasseron C, et al. Development and Validation of a Risk Score Predicting Death Without Transplant in Adult Heart Transplant Candidates. JAMA. 2024;331(6):500–509. doi:10.1001/jama.2023.27029
  2. Kittleson MM. Optimizing Beneficence and Justice in Heart Transplant Allocation. JAMA. 2024;331(6):480–481. doi:10.1001/jama.2023.27157