MELD-based liver allocation has been widely heralded for its ability to accurately prioritize patients according to waitlist mortality (1,2). This policy has resulted in more equitable racial access to liver transplantation (2). But, as a growing body of evidence has demonstrated, gender disparity exists.
Previous studies using United Network for Organ Sharing registry data have reported that women experience lower transplantation rates than men (2,3). In this month’s AJT, Mathur et al. expand upon our understanding of this gender disparity (4). Analyses adjusted for the donation service area (DSA), a factor that strongly influences organ acceptance practices, confirm this finding. Moreover, the disparity has actually widened from the pre-MELD (1997–2002) to the MELD (2002–2007) era. The cumulative weight of these multiple reports compels us to move beyond challenging whether gender disparity exists to explaining why.
Two reasons have been commonly put forward to explain the gender disparity in liver transplant rates. The first is that inclusion of serum creatinine in the MELD score results in under-representation of liver disease severity in women relative to men. This is a nonviable explanation as analyses reflect transplant rates at a given MELD score (2–4). The second is that female candidates, generally smaller than males, wait longer for a size-appropriate graft from the donor pool that is predominantly male. Again, proper statistical analyses that not only adjusted for body size, but also explored the possible interaction between gender and height (which was not significant), do not fully support this theory (3,4).
Instead, we propose alternative hypotheses. Female and male liver transplant candidates differ significantly by liver disease diagnosis. Different diseases may be characterized by different rates of change in the MELD score over time (ΔMELD). For example, at the same MELD score, a woman with primary biliary cirrhosis with a lower ΔMELD may be perceived as less sick than a man with hepatitis C with a higher ΔMELD. Previous studies have shown that ΔMELD provides important prognostic information above and beyond the MELD score at a single point in time (5). However, it is unknown whether this metric captures disease severity equally between women and men. Providers, weighing the risk of wait-list mortality between a female and a male candidate with the same MELD score but a different ΔMELD, may (subconsciously) favor transplantation of the candidate with the greatest ΔMELD, potentially leading to the gender bias in transplantation rates. Well-designed studies to better characterize the association between MELD score, ΔMELD and waitlist mortality by gender are eminently feasible and greatly needed.
More challenging to evaluate, however, is the impact of noncandidate factors, such as geographic differences in liver distribution, the hepatocellular carcinoma (HCC) exception policy and provider perception of disease severity on accentuating the gender disparity after implementation of MELD-based allocation. The striking geographic variation in liver transplantation rates by gender in the MELD era (4) likely reflects a multitude of factors including distinct center practices, competition among centers within a DSA or regional differences in organ availability. Alternatively, although patients receiving HCC exception points were excluded from this study (4), it is possible that the increase in transplants for HCC patients, who are predominantly male, decreases the availability of organs to non-HCC candidates, exerting a disproportionate effect on females in the MELD era. In addition, providers may unintentionally contribute to gender disparity through disparate organ acceptance practices for female versus male candidates.
Which of these factors, or combination of factors, has led to the widening gender disparity in liver transplant rates is unknown. But perhaps the most compelling question is: Do we even care about differences in transplantation rates if there are no differences in waitlist mortality (4)? This would suggest that the MELD-based allocation system is, in fact, doing precisely what it was designed to do. Perhaps women have a lower risk of death than men with the same MELD score and can therefore ‘afford’ to wait longer. Might then differences in transplant rates simply reflect gender differences in disease progression and risk of death at a given MELD score? This issue is controversial, but one that can and must be resolved.
Mathur’s article is an important demonstration that, 9 years after the implementation of MELD-based allocation, fundamental questions regarding equity remain. These questions may be rooted in an insufficient understanding of the complexities of gender differences related to various liver diseases and their progression. In addition, sociopolitical factors likely play a key role but are elusive and difficult to define. It is certainly sobering to realize that, while MELD allocation is a significant step forward, we still have a long way to go before arriving at a fully equitable policy.