AimDifferential patterns of peripheral memory T cell subsets in non‐alcoholic fatty liver disease (NAFLD) were assessed using flow cytometry (FCM) to elucidate their association with NAFLD severity and provide a new non‐invasive method to sensitively detect the disease severity in addition to existing biomarkers.MethodsWe assessed the differential frequencies of peripheral memory T cell subsets in 103 patients with NAFLD according to the degree of liver fibrosis via FCM analysis. We focused on the following populations: CCR7+ CD45RA+ naïve T, CCR7+ CD45RA– central memory T (TCM), CCR7– CD45RA– effector memory T, and CCR7– CD45RA+ terminally differentiated effector memory T (TEMRA) cells in CD4+ and CD8+ T, Th1, Th2, and Th17 cells, respectively. To evaluate the pathological progression of the disease, these frequencies were also examined according to the degree of the NAFLD activity score (NAS).ResultsSeveral significant correlations were observed between laboratory parameters and peripheral memory T lymphocyte frequencies according to the degree of liver fibrosis and NAS in NAFLD. In univariate and multivariate analyses, the frequency of CD8+ TEMRA cells predicted severe fibrosis, and the predictive power was validated in an independent cohort. Furthermore, the frequencies of several memory T cell subsets sensitively indicated the pathological progression of NAFLD (Th17 TCM: steatosis, CD4+ TCM: lobular inflammation, and CD8+ TEMRA and TEM: hepatocellular ballooning).ConclusionsOur results suggest that the analysis of peripheral memory T lymphocyte frequencies can non‐invasively predict severe fibrosis and sensitively indicate the pathological progression of NAFLD.This article is protected by copyright. All rights reserved.