Use of serum sodium for liver transplant graft allocation: A decade in the making, now is it ready for primetime?

See Article on Page 308 Hyponatremia has long been known to be a harbinger of death in cirrhosis. In 1956, Dame Sheila Sherlock observed that “in patients with liver disease, serum sodium levels below 130 mEq/L must be regarded as serious and, if below 125 mEq/L, ominous.”1 This astute clinical observation was substantiated in several subsequent studies.2 Over the last decade, serum sodium has been extensively investigated specifically as an adjunct to the Model for End‐Stage Liver Disease (MELD) score in predicting wait‐list death and the urgency for liver transplantation.8 In patients waiting for liver transplantation and particularly in patients with cirrhosis and a low MELD score, hyponatremia is a strong independent predictor of wait‐list death. Serum sodium has been incorporated into MELD in a manner that accounts for the larger prognostic influence when the MELD score is low and for a smaller effect when the MELD score is high.8 The MELD‐Na score, which incorporates sodium into the MELD score, was approved in June 2014 by the United Network for Organ Sharing (UNOS)/Organ Procurement and Transplantation Network (OPTN) board of directors for implementation as the measure of urgency for liver transplantation in the United States.15 In this issue of Liver Transplantation, Sharma et al. investigate the influence that serum sodium has on the survival benefit achieved from liver transplantation. In a previous landmark study from the same group, liver transplant candidates with MELD scores < 15 were shown to not achieve a survival benefit from liver transplantation.17 The current study by Sharma et al. answers 2 important questions with respect to implementing MELD‐Na for liver allocation: Will patients with a low MELD score and a low serum sodium level achieve a survival benefit from liver transplantation? Is there a lower threshold for the MELD score at which liver transplant candidates do not get a survival benefit despite the presence of hyponatremia? The survival benefit of liver transplantation is the ratio of the risk of death after transplantation to the risk of death while one is waiting for liver transplantation. As the risk of death on the wait‐list increases, the denominator of this ratio increases, and as a result, the likelihood of achieving a survival benefit increases as long as the numerator (the risk of death after transplantation) does not similarly increase. Although there may be an increase in short‐term complications (neurologic disorders, infections, and renal failure) after liver transplantation in recipients with pretransplant hyponatremia,18 the influence on posttransplant mortality has been shown to be undetectable14 or, at most, small18 in comparison with the increase in the risk of death without liver transplantation. Therefore, it is not surprising that Sharma et al.16 found that for most liver transplant candidates, there is an increase in the survival benefit from a transplant with decreasing serum sodium, and this effect is driven by increased wait‐list mortality. The important exception identified by Sharma et al. is that patients with MELD scores ≤ 11 were found to not achieve a survival benefit whether they had hyponatremia or not. For this reason, the UNOS/OPTN board of directors has chosen to limit the application of MELD‐Na to liver transplant candidates with MELD scores > 11.15 The Scientific Registry of Transplant Recipients (SRTR) and the UNOS Liver and Intestine Committee examined several ways to incorporate serum sodium into liver allocation in the United States, and they settled on the MELD‐Na equation22: (1)MELD‐Na = MELD + 1.32 × (137 – Na) − [0.033 × MELD × (137 – Na)] Modeling performed by the SRTR, with a lower limit of sodium of 125 mEq/L and an upper limit of 137 mEq/L, predicted that in comparison with MELD, the use of MELD‐Na for allocation would result in 52 fewer total (wait‐list and posttransplant) deaths per year.22 This analysis did not exclude patients with pretransplant MELD scores ≤ 11 yet because liver transplantation is rare at such a low MELD score, and the influence of including patients with MELD scores ≤ 11 is likely to be quite small. So, after a decade of considering the use of serum sodium in a liver allocation tool, are we ready for its implementation? Critics of such a policy change have raised concerns about the variability of serum sodium over time; proponents counter that the mandatory scheduled updates of MELD laboratory tests will be sufficient to blunt the influence of such variability. Critics worry about creating worse posttransplant outcomes by performing transplantation for sicker patients; proponents reply that this same concern existed with the implementation of MELD but did not occur. Critics predict that providers or patients may modify their clinical management to intentionally lower serum sodium and thus raise the MELD‐Na score; proponents rebut that the other parameters in the MELD score are subject to similar pressures. In the end, what is clear is that accurately modeling the behavioral response to a policy change is difficult, if not impossible. There will likely be some unintended consequences with the implementation of MELD‐Na as a liver allocation tool. Postimplementation evaluation of MELD‐Na will be imperative to characterize the behavioral responses and to assess unintended consequences. Several observational studies,8 a regional pilot assessment,23 and extensive predictive modeling,22 thus far, have reliably shown that a relatively inexpensive computer coding change can improve the prediction of the urgency for liver transplantation in our “sickest first” policy and can save lives. After a decade of consideration of MELD‐Na, further pre‐implementation analysis is unlikely to further inform policy decision making but certainly would delay a policy change that may save more than 50 lives per year.