Drug & Device Development

New blood test detects non-alcoholic steatohepatitis

Last Updated: 2006-07-13 13:35:31 -0400 (Reuters Health)

NEW YORK (Reuters Health) - Measuring plasma levels of cytokeratin-18 fragments (CK-18) is a noninvasive way to differentiate nonalcoholic steatohepatitis (NASH) from simple steatosis or normal tissue, according to investigators at the Cleveland Clinic in Ohio.

Hepatocyte apoptosis characterizes NASH, and in late stages of the disease, activated caspases cleave CK-18, the major intermediate filament protein in the liver. Accurate diagnosis NASH is important, the research team notes, because about one quarter of these patients progress to cirrhosis, liver failure or hepatocellular carcinoma.

However, up until now, histologic examination of a liver biopsy has been the only means of diagnosing NASH and assessing its severity.

To determine whether measurement of caspace-3 generated CK-18 was accurate enough to substitute for a liver biopsy, Dr. Ariel E. Feldstein's and his team compared the two tests among 39 patients with clinically suspected nonalcoholic fatty liver disease. They report their findings in a rapid communication in the July issue of Hepatology.

Biopsy results showed that 8 had simple steatosis, 21 had definitive NASH, and 10 were considered to be normal. Cirrhosis was diagnosed in 2 patients with definite NASH.

Median plasma CK-18 levels of all 39 patients were significantly higher than that of 35 healthy age-matched controls (516.7 U/L versus 234, p < 0.001), the investigators report.

CK-18 levels were "strikingly higher" in patients with definitive NASH than in patients with normal biopsies or simple steatosis (766, 212 and 215 U/L, respectively, p < 0.001).

The researchers note that CK-18 fragment levels did not correlate with age, serum transaminase levels, diabetes, dyslipidemia or hypertension. There were weak correlations with body mass index and stage of fibrosis.

The odds ratio of having definitive NASH remained significant (1.95) after adjusting for these variables, "confirming that elevated levels of CK-18 fragments served as an independent predictor of 'definitive NASH' (p < 0.05)," Dr. Feldstein's team reports.

The authors found that using a cutoff value of 395 U/L to minimize the rate of false positives, resulted in a sensitivity of 85.7% and specificity of 99.9%. To minimize false negatives, a cutoff value of 380.2 U/L had a sensitivity of 90.5% and specificity of 94.4%.

Dr. Feldstein's team concludes that this noninvasive test may "aid clinicians in the selection of patients for liver biopsy (and) also allow for noninvasive assessment of disease progression and therapeutic response."

Hepatology 2006;44:27-33.