Creatinine variability linked to liver transplant outcomes

Patients with greater changes in serum creatinine are more likely to have worse pre- and post–liver transplant outcomes. Moreover, underserved patients may be most frequently affected, according to a retrospective analysis of UNOS (United Network for Organ Sharing) data.

These results should drive further development of serum creatinine coefficient of variation (sCr CoV) as an independent predictor of renal-related mortality risk, according to lead author Giuseppe Cullaro, MD, of the University of California, San Francisco, and colleagues.

“Intra-individual clinical and laboratory parameter dynamics often provide additional prognostic information – added information that goes beyond what can be found with cross-sectional data,” the researchers wrote in Hepatology. “This finding has been seen in several scenarios in the general population – intra-individual variability in blood pressure, weight, hemoglobin, and kidney function, have all been associated with worse clinical outcomes. However, in cirrhosis patients, and more specifically in patients awaiting a liver transplant, kidney function dynamics as a predictor of clinical outcomes has yet to be investigated.”

To gauge the predictive power of shifting kidney values, Dr. Cullaro and colleagues analyzed UNOS/OPTN (Organ Procurement and Transplantation Network) registry data from 2011 through 2019. Exclusion criteria included patients who were aged younger than 18 years, were listed as status 1, received a living donor liver transplantation, were on hemodialysis, or had fewer than three updates. The final dataset included 25,204 patients.

After the researchers sorted patients into low, intermediate, and high sCr CoV tertiles, they used logistic regression to determine relationships between higher sCr and a variety of covariates, such as age, sex, diagnosis, presence of acute kidney injury, or chronic kidney disease. A competing risk regression was then done to look for associations between wait list mortality and the covariables, with liver transplant used as the competing risk.

The median sCr CoV was 17.4% (interquartile range [IQR], 10.8%-29.5%). Patients in the bottom sCr CoV tertile had a median value of 8.8% (IQR, 6.6%-10.8%), compared with 17.4% (IQR, 14.8%-20.4%) in the intermediate variability group and 36.8% (IQR, 29.5%-48.8%) in the high variability group. High variability was associated with female sex, Hispanic ethnicity, ascites, and hepatic encephalopathy as well as higher body mass index, MELDNa score, and serum creatinine.

Of note, each decreasing serum creatinine variability tertile was associated with a significantly lower rate of wait list mortality (34.7% vs. 19.6% vs. 11.7%; P < .001). The creatinine variability profiles were similarly associated with the likelihood of receiving a liver transplant (52.3% vs. 48.9% vs. 43.7%; P < .001) and posttransplant mortality (7.5% vs. 5.5% vs. 3.9%; P < .001).

A multivariate model showed that each 10% increase in sCr CoV predicted an 8% increased risk of a combined outcome comprising post–liver transplant death or post–liver transplant kidney transplant (KALT), independently of other variables (adjusted hazard ratio, 1.08; 95% confidence interval, 1.05-1.11).

“These data highlight that all fluctuations in sCr are associated with worse pre- and post–liver transplant outcomes,” the investigators concluded. “Moreover, the groups that are most underserved by sCr, specifically women, were most likely to have greater sCr CoVs. We believe our work lays the foundation for implementing the sCr CoV as an independent metric of renal-related mortality risk and may be most beneficial for those groups most underserved by sCr values alone.”

According to Brian P. Lee, MD, a hepatologist with Keck Medicine of USC and assistant professor of clinical medicine with the Keck School of Medicine of USC in Los Angeles, “this is a great study ... in an area of high need” that used “high quality data.”

Current liver allocation strategies depend on a snapshot of kidney function, but these new findings suggest that a more dynamic approach may be needed. “As a practicing liver specialist I see that creatinine numbers can fluctuate a lot. ... So which number do you use when you’re trying to calculate what a patient’s risk of death is on the wait list? This study gets toward that answer. If there is a lot of variability, these might be higher risk patients; these might be patients that we should put higher on the transplant waiting list,” said Dr. Lee.

He suggested that clinicians should account for creatinine fluctuations when considering mortality risk; however, the evidence is “not quite there yet ... in terms of changing transplant policy and allocation.” He pointed out three unanswered questions: Why are creatinine values fluctuating? How should fluctuations be scored for risk modeling? And, what impact would those risk scores have on transplant waitlist prioritization?

“I think that that’s the work that you would need to do before you could really change national transplant policy,” Dr. Lee concluded.

The study was supported by the National Institutes of Health and the UCSF Liver Center. Dr. Cullaro and another author have disclosed relationships with Mallinckrodt Pharmaceuticals and Axcella Health, respectively. Dr. Lee reported no conflicts of interest.

Patients with greater changes in serum creatinine are more likely to have worse pre- and post–liver transplant outcomes. Moreover, underserved patients may be most frequently affected, according to a retrospective analysis of UNOS (United Network for Organ Sharing) data.

These results should drive further development of serum creatinine coefficient of variation (sCr CoV) as an independent predictor of renal-related mortality risk, according to lead author Giuseppe Cullaro, MD, of the University of California, San Francisco, and colleagues.

“Intra-individual clinical and laboratory parameter dynamics often provide additional prognostic information – added information that goes beyond what can be found with cross-sectional data,” the researchers wrote in Hepatology. “This finding has been seen in several scenarios in the general population – intra-individual variability in blood pressure, weight, hemoglobin, and kidney function, have all been associated with worse clinical outcomes. However, in cirrhosis patients, and more specifically in patients awaiting a liver transplant, kidney function dynamics as a predictor of clinical outcomes has yet to be investigated.”

To gauge the predictive power of shifting kidney values, Dr. Cullaro and colleagues analyzed UNOS/OPTN (Organ Procurement and Transplantation Network) registry data from 2011 through 2019. Exclusion criteria included patients who were aged younger than 18 years, were listed as status 1, received a living donor liver transplantation, were on hemodialysis, or had fewer than three updates. The final dataset included 25,204 patients.

After the researchers sorted patients into low, intermediate, and high sCr CoV tertiles, they used logistic regression to determine relationships between higher sCr and a variety of covariates, such as age, sex, diagnosis, presence of acute kidney injury, or chronic kidney disease. A competing risk regression was then done to look for associations between wait list mortality and the covariables, with liver transplant used as the competing risk.

The median sCr CoV was 17.4% (interquartile range [IQR], 10.8%-29.5%). Patients in the bottom sCr CoV tertile had a median value of 8.8% (IQR, 6.6%-10.8%), compared with 17.4% (IQR, 14.8%-20.4%) in the intermediate variability group and 36.8% (IQR, 29.5%-48.8%) in the high variability group. High variability was associated with female sex, Hispanic ethnicity, ascites, and hepatic encephalopathy as well as higher body mass index, MELDNa score, and serum creatinine.

Of note, each decreasing serum creatinine variability tertile was associated with a significantly lower rate of wait list mortality (34.7% vs. 19.6% vs. 11.7%; P < .001). The creatinine variability profiles were similarly associated with the likelihood of receiving a liver transplant (52.3% vs. 48.9% vs. 43.7%; P < .001) and posttransplant mortality (7.5% vs. 5.5% vs. 3.9%; P < .001).

A multivariate model showed that each 10% increase in sCr CoV predicted an 8% increased risk of a combined outcome comprising post–liver transplant death or post–liver transplant kidney transplant (KALT), independently of other variables (adjusted hazard ratio, 1.08; 95% confidence interval, 1.05-1.11).

“These data highlight that all fluctuations in sCr are associated with worse pre- and post–liver transplant outcomes,” the investigators concluded. “Moreover, the groups that are most underserved by sCr, specifically women, were most likely to have greater sCr CoVs. We believe our work lays the foundation for implementing the sCr CoV as an independent metric of renal-related mortality risk and may be most beneficial for those groups most underserved by sCr values alone.”

According to Brian P. Lee, MD, a hepatologist with Keck Medicine of USC and assistant professor of clinical medicine with the Keck School of Medicine of USC in Los Angeles, “this is a great study ... in an area of high need” that used “high quality data.”

Current liver allocation strategies depend on a snapshot of kidney function, but these new findings suggest that a more dynamic approach may be needed. “As a practicing liver specialist I see that creatinine numbers can fluctuate a lot. ... So which number do you use when you’re trying to calculate what a patient’s risk of death is on the wait list? This study gets toward that answer. If there is a lot of variability, these might be higher risk patients; these might be patients that we should put higher on the transplant waiting list,” said Dr. Lee.

He suggested that clinicians should account for creatinine fluctuations when considering mortality risk; however, the evidence is “not quite there yet ... in terms of changing transplant policy and allocation.” He pointed out three unanswered questions: Why are creatinine values fluctuating? How should fluctuations be scored for risk modeling? And, what impact would those risk scores have on transplant waitlist prioritization?

“I think that that’s the work that you would need to do before you could really change national transplant policy,” Dr. Lee concluded.

The study was supported by the National Institutes of Health and the UCSF Liver Center. Dr. Cullaro and another author have disclosed relationships with Mallinckrodt Pharmaceuticals and Axcella Health, respectively. Dr. Lee reported no conflicts of interest.

Patients with greater changes in serum creatinine are more likely to have worse pre- and post–liver transplant outcomes. Moreover, underserved patients may be most frequently affected, according to a retrospective analysis of UNOS (United Network for Organ Sharing) data.

These results should drive further development of serum creatinine coefficient of variation (sCr CoV) as an independent predictor of renal-related mortality risk, according to lead author Giuseppe Cullaro, MD, of the University of California, San Francisco, and colleagues.

“Intra-individual clinical and laboratory parameter dynamics often provide additional prognostic information – added information that goes beyond what can be found with cross-sectional data,” the researchers wrote in Hepatology. “This finding has been seen in several scenarios in the general population – intra-individual variability in blood pressure, weight, hemoglobin, and kidney function, have all been associated with worse clinical outcomes. However, in cirrhosis patients, and more specifically in patients awaiting a liver transplant, kidney function dynamics as a predictor of clinical outcomes has yet to be investigated.”

To gauge the predictive power of shifting kidney values, Dr. Cullaro and colleagues analyzed UNOS/OPTN (Organ Procurement and Transplantation Network) registry data from 2011 through 2019. Exclusion criteria included patients who were aged younger than 18 years, were listed as status 1, received a living donor liver transplantation, were on hemodialysis, or had fewer than three updates. The final dataset included 25,204 patients.

After the researchers sorted patients into low, intermediate, and high sCr CoV tertiles, they used logistic regression to determine relationships between higher sCr and a variety of covariates, such as age, sex, diagnosis, presence of acute kidney injury, or chronic kidney disease. A competing risk regression was then done to look for associations between wait list mortality and the covariables, with liver transplant used as the competing risk.

The median sCr CoV was 17.4% (interquartile range [IQR], 10.8%-29.5%). Patients in the bottom sCr CoV tertile had a median value of 8.8% (IQR, 6.6%-10.8%), compared with 17.4% (IQR, 14.8%-20.4%) in the intermediate variability group and 36.8% (IQR, 29.5%-48.8%) in the high variability group. High variability was associated with female sex, Hispanic ethnicity, ascites, and hepatic encephalopathy as well as higher body mass index, MELDNa score, and serum creatinine.

Of note, each decreasing serum creatinine variability tertile was associated with a significantly lower rate of wait list mortality (34.7% vs. 19.6% vs. 11.7%; P < .001). The creatinine variability profiles were similarly associated with the likelihood of receiving a liver transplant (52.3% vs. 48.9% vs. 43.7%; P < .001) and posttransplant mortality (7.5% vs. 5.5% vs. 3.9%; P < .001).

A multivariate model showed that each 10% increase in sCr CoV predicted an 8% increased risk of a combined outcome comprising post–liver transplant death or post–liver transplant kidney transplant (KALT), independently of other variables (adjusted hazard ratio, 1.08; 95% confidence interval, 1.05-1.11).

“These data highlight that all fluctuations in sCr are associated with worse pre- and post–liver transplant outcomes,” the investigators concluded. “Moreover, the groups that are most underserved by sCr, specifically women, were most likely to have greater sCr CoVs. We believe our work lays the foundation for implementing the sCr CoV as an independent metric of renal-related mortality risk and may be most beneficial for those groups most underserved by sCr values alone.”

According to Brian P. Lee, MD, a hepatologist with Keck Medicine of USC and assistant professor of clinical medicine with the Keck School of Medicine of USC in Los Angeles, “this is a great study ... in an area of high need” that used “high quality data.”

Current liver allocation strategies depend on a snapshot of kidney function, but these new findings suggest that a more dynamic approach may be needed. “As a practicing liver specialist I see that creatinine numbers can fluctuate a lot. ... So which number do you use when you’re trying to calculate what a patient’s risk of death is on the wait list? This study gets toward that answer. If there is a lot of variability, these might be higher risk patients; these might be patients that we should put higher on the transplant waiting list,” said Dr. Lee.

He suggested that clinicians should account for creatinine fluctuations when considering mortality risk; however, the evidence is “not quite there yet ... in terms of changing transplant policy and allocation.” He pointed out three unanswered questions: Why are creatinine values fluctuating? How should fluctuations be scored for risk modeling? And, what impact would those risk scores have on transplant waitlist prioritization?

“I think that that’s the work that you would need to do before you could really change national transplant policy,” Dr. Lee concluded.

The study was supported by the National Institutes of Health and the UCSF Liver Center. Dr. Cullaro and another author have disclosed relationships with Mallinckrodt Pharmaceuticals and Axcella Health, respectively. Dr. Lee reported no conflicts of interest.