This article is more than 5 years old. Information may no longer be current.

MELD-GRAIL-Na score may improve liver transplant waitlist mortality prediction

Replacing serum creatinine with estimated glomerular filtration rate in a patient’s MELD-Na score appeared to improve waitlist mortality prediction among patients with cirrhosis in need of a liver transplantation, according to study results published in Hepatology.

“The MELD and MELD-Na are excellent predictors of [waitlist] mortality. However, as compared to when these were developed over a decade ago, the transplant population has changed,” Sumeet K. Asrani, MD, MSc, from the Baylor University Medical Center in Texas, and colleagues wrote. “Accurate prediction of mortality among patients awaiting transplantation is important as it forms the backbone of urgency-based organ allocation.”

The changes that the researchers referred to include increasing rates of MELD scores of 30 or higher (0.4% vs. 3.2%), more patients aged 65 years or older (12% vs. 24%), higher rates of diabetes (23.5% vs. 28.8%), and lower rates of viral hepatitis as the indication for liver transplantation (25.2% vs. 17.6%).
To examine 90-day waitlist mortality prediction methods, Asrani and colleagues compared MELD-Na with a combination of re-estimated bilirubin, international normalized ratio, sodium and glomerular filtration rate assessment in liver disease, or MELD-GRAIL-Na, among 17,095 patients awaiting a liver transplant for cirrhosis.

While the performance of MELD and MELD-Na remained excellent, it worsened over time from 2005 to 2015 and varied by disease severity.

In contrast, MELD-GRAIL (HR = 1.209; 95% CI, 1.197-1.221) and MELD-GRAIL-Na (HR = 1.212; 95% CI, 1.199-1.224) demonstrated significant waitlist mortality prediction. MELD-GRAIL-Na also had better discrimination (c = 0.83; P < .001) compared with MELD (c = 0.81) and MELD-Na (c = 0.82).

Results from subgroup analysis revealed that MELD-GRAIL-Na had better discrimination for women older than 60 years (c = 0.81 vs. 0.8; P < .001) and for women older than 60 with MELD higher than 35 (c = 0.58 vs. 0.53; P < .01) compared with MELD.

“We estimate that incorporation of these models may impact outcomes for 12% to 17% of registrants awaiting transplant,” the researchers concluded. “MELD-GRAIL-Na may serve as an improvement in the model for prediction of WL mortality and organ allocation policy compared to current state. The ability to further improve performance in the older, underweight and sicker women may also be beneficial.” – by Talitha Bennett

Disclosures: The authors report no relevant financial disclosures.