Background: Previous cross-sectional studies have shown that serum levels of cytokeratin fragments (CK18) are associated with the presence of steatohepatitis in individuals with nonalcoholic fatty liver disease (NAFLD). However, it is unknown if serial serum CK18 levels can predict longitudinal changes in liver histology in NAFLD. Aim: To determine the degree to which changes in serum CK18 associate with changes in liver histology over a 96-week period in adult patients with nonalcoholic steatohepatitis (NASH). Methods: Serum CK18 levels were measured at baseline and 16, 48, and 96 weeks in 231 of the 247 adults with NASH who participated in the PIVENS trial which investigated the efficacy of pioglitazone vs. vitamin E vs. placebo in non-diabetic individuals with histologically confirmed NASH over a 96-week period. Liver biopsies at baseline and after 96 weeks of treatment were centrally evaluated by the NASH CRN Pathology Committee. Multiple logistic (for categorical outcomes) and linear (for continuous outcome variables) regression models were used to determine the association between changes in histological features, ALT, and % collagen assessed by Sirius red staining, and change in CK18 levels, controlling for baseline CK18 levels and treatment group. Results: At baseline, CK18 levels among the 3 treatment groups were similar (placebo: 440± 350 U/L, vit E: 510± 350 U/L, and pioglitazone: 490± 410 U/L). Compared to placebo, serum CK18 levels among individuals treated with vit E were reduced at week 16 (mean change from baseline -160± 300 vs. -50 ±380 U/L, p= 0.02), week 48 (-220± 390 vs. -10± 370 U/L, p=0.009), and week 96 (-200± 400 vs. 30± 400 U/L, p=0.009). Similarly, compared to placebo, serum CK18 levels among individuals treated with pioglitazone were reduced at week 16 (mean change from baseline -240 ± 390 vs. -50± 380 U/L, p,=0.001), week 48 (-180± 370 vs. -10± 370 U/L, p=0.001, and week 96 (-260 ±400 vs. 30 ± 400 U/L, p=0.001). Strong correlations were seen with change in CK18 levels and the primary histologic endpoint, resolution of NASH, and individual histologic features of NASH (Table 1). In addition, change in CK18 levels strongly correlated with change in ALT levels (b=0.06, 95% CI: 0.04, 0.08; P,0.001).Percent collagen measurements were not related to change in CK18 (b=-0.17 %/100*U/L , 95% CI:-0.70, 0.37; P= 0.54). Summary: (1) Compared to placebo, treatment with vit E or pioglitazone had significant reduction in serum CK18 levels. (2) Changes in serum CK18 correlated with histologic improvement in non-diabetic adults with NASH treated with vitamin E or pioglitazone. Conclusion: Serum CK18 is a potential useful surrogate marker for detection of improvement in clinical trials.