DDW: Blood Test May Replace Biopsy In Fatty Liver Diagnosis

When fatty liver is the problem, a simple blood test can determine as well as an hepatic biopsy whether the condition is dangerous or benign, a researcher said here. In a large cohort study, the blood test -- for a marker of liver cell death -- was highly accurate in distinguishing non-alcoholic steatohepatitis (NASH) from other forms of fatty liver and from healthy livers, found Ariel Feldstein, M.D., of the Cleveland Clinic.
The test looks for fragments of cytokeratin-18, a liver filament protein that is broken down specifically during liver cell death, Dr. Feldstein told a session at Digestive Disease Week.
The level of cytokeratin-18 fragments is significantly elevated (P<0.001) in patients with NASH, compared with those who have simple steatosis or a healthy liver, Dr. Feldstein said.
He and colleagues had previously shown that cell death plays a role in NASH, but not in patients with simple steatotis, which is regarded as a benign condition that does not progress.
To follow up, they enrolled 178 patients (participants in several fatty liver clinical trials) whose condition was well characterized by biopsy. Also, they enrolled 150 age-matched healthy controls.
Using a monoclonal antibody to cytokeratin-18 fragments, they found that the plasma level of the fragments predicted the likelihood of having NASH, as opposed to simple steatosis.
Specifically, every increase of 50 U/L increased the likelihood of NASH (opposed to steatosis) by 74%, which was significant at P=0.0001. The odds ratio was 1.74, with a 95% confidence interval from 1.31 to 2.31.
Among patients with any form of fatty liver, the plasma level of the fragments ranged from 68 to 3000 U/L, with a median of 244. Patients with benign fatty liver had a median of 182 and those with borderline disease had a level of 179. The healthy controls had a median level of 145 U/L.
By contrast, those with NASH had a median level of 361, Dr. Feldstein said.
If the cutoff for NASH is set at 246 U/L, he said, the test predicts the disease with a specificity of 81% and a sensitivity of 75%.
The test for the cytokeratin-18 fragments is commercially available, Dr. Feldstein said, but only as a research tool, and it's not likely to be available for clinical use for a couple of years.
The test is "certainly farther along than other biomarkers we've seen for liver disease," commented Jacquelyn Maher, M.D., of the University of California San Francisco.
But "there are still some bugs to be worked out," said Dr. Maher, who was not involved in the research. Specifically, she said, "there's not a lot of wiggle room" between those with frank NASH and those with borderline disease, which may limit diagnostic utility.