Acute-on-chronic liver failure model accurately predicts 30-day survival
Investigators have validated the North American Consortium for the Study of End-Stage Liver Disease Acute-On Chronic Liver Failure model, or NACSELD-ACLF, as an accurate tool to predict 30-day survival in a study of patients with or without acute infection and cirrhosis.
“NACSELD is a multi-center North American Consortium focused on cirrhosis outcomes using prospectively collected data,” Jasmohan S. Bajaj, MD, from the Virginia Commonwealth University, told Healio Gastroenterology and Liver Disease. “The NACSELD-ACLF was developed first for infections as reported in our publication in 2014 in Hepatology in order to provide a very simple method to gauge negative outcomes in inpatients with cirrhosis.
“This simple score of two or more features of organ failures including dialysis, grade 3/4 hepatic encephalopathy, shock or mechanical ventilation was developed from this initial report with deaths as outcome,” Bajaj continued. “We then wanted to validate it in a larger separate cohort and also in patients who did not have infections. This is what the most recent paper does.”
Bajaj and colleagues prospectively analyzed the data of 2,675 patients admitted to one of 14 NACSELD sites, 1,079 of whom had an acute infection at admission or developed an acute infection during their initial hospital stay.
Patients with acute infection were more likely to be women (40% vs. 36%; P = .04), had higher heart rates (86.5 vs. 84.36; P = .003), temperatures (98.17 vs. 97.95; P = .0001), white blood cell counts (5.7 vs 4.8; P = .002) and serum creatinine levels (1.48 vs. 1.39; P < .0001), and lower systolic blood pressure (119.83 vs. 122.63; P = .0009), serum albumin levels (2.75 vs. 2.91; P < .0001) and serum sodium levels (133.57 vs. 134.71; P < .0001) compared with uninfected patients.
Additionally, crude 30-day survival rates were lower in patients with acute infection (84% vs. 93%; P < .0001) and those with organ failures.
Patients who met NACSELD-ACLF criteria of two or more organ failures had an overall 30-day survival of 59% compared with 93% in those without NACSELD-ACLF. The 30-day survival rates among patients who met NACSELD-ACLF criteria and had infection was 52% compared with 76% in those who met NACSELF-ACLF but did not have infection.
The strongest predictor of 30-day survival was NACSELD-ACLF among both patients with acute infection (OR = 0.16; 95% CI, 0.11-0.32) and those without infection (OR = 0.29; 95% CI, 0.15-0.56) compared with a previous model and after controlling for MELD score, white blood cell count and admission serum albumin.
Bajaj and colleagues also compared the NACSELD-ACLF model with the Asian-Pacific Association for the Study of the Liver (APASL) model to predict patient survival. To assess the NACSELD-ACLF model, the researchers eliminated patients admitted with infections (n = 741) and those who were uninfected but developed NACSELD-ACLF within 48 hours of admission (n = 59). Similarly, to assess the APASL model, the researchers eliminated patients admitted with an infection and those who had serum bilirubin higher than 5 mg/dL or international normalized ratio higher than 1.5 at admission (n = 681).
In those new cohorts, 30-day survival rates were 55% among patients who met NACSELD-ACLF criteria vs. 93% among those who did not and 87% among those who met APASL criteria vs. 95% among those who did not (P < .001).
“There are several other mechanisms that can be used for inpatients who are critical such as APACHE, SOFA, CLIF-SOFA, etc., but some of them are quite complicated. Any score that is complicated adds one more layer of difficulty that clinicians have to fight against in their clinical practice,” Bajaj concluded. “Therefore, just looking at the patient or even their chart during the hospitalization to gauge the four organ failures noted above is easy and is actually standard of care.” – by Talitha Bennett
Disclosure: The authors report relevant financial disclosures.