Poor Inter-test Reliability Between CK18 Kits as a Biomarker of NASH.

Abstract

Nonalcoholic fatty liver disease (NAFLD) affects 15-40% of the general population; 10-20% of those patients have a more severe form of the disease known as nonalcoholic steatohepatitis (NASH). Cytokeratin-18 (CK18), released during apoptosis and one of the most studied biomarkers in NASH, can be measured by a number of commercially available kits. We compared serum measurements of the CK18 M30 from two different kits using the same cohort to evaluate the reliability between two test kits. We measured serum levels of CK18 M30 from 185 patients with biopsy-proven NAFLD from a single center from 2009 to 2015, using two different ELISA kits, Test 1 (T1) and Test 2 (T2). Advanced fibrosis was defined as fibrosis stages 3-4 and NASH defined by NAS score ≥ 5. Mean age was 50.2 years (SD 12.6), 61.1% male and 87% White; 49.6% had NASH and 32.2% advanced fibrosis. There was no significant correlation between measurements from the two kits (p = 0.86, r = 0.01). While T2 predicted NASH and advanced fibrosis, T1 did not. The area under ROC curve for the prediction of NASH was 0.631 for T2 versus 0.500 for T1. Measurements from two different CK18 M30 test kits did not correlate with each other. One kit showed statistically significantly higher levels of CK18 M30 in patients with advanced fibrosis and NASH, while the other kit did not. With the increasing use of CK18 as a biomarker in NASH, it is important to standardize the different kits as it could greatly bias the results.