Redefining the upper limit or normal (ULN) of alanine transaminase (ALT) levels for Asians: a systematic review and meta-analysis of 60 studies and 335,163 individuals

Abstract

Background: Current ULNs of ALT were defined using older data that did not specifically exclude patients with non‐alcoholic fatty liver disease (NAFLD). With the obesity epidemic, the ULN for ALT may have changed. Asians also have lower body mass, which can affect ALT levels. Therefore, we aimed to estimate the mean and ULN of ALT in Asians without liver disease including NAFLD and viral hepatitis using a systematic review and meta‐analytic approach with literature from Pubmed, Embase and Cochrane databases up to January 17, 2019. We also obtained individual patient level data from some of the study cohorts to validate study level data. Methods: We initially searched for studies investigating NAFLD to identify study cohorts that were screened for viral hepatitis, alcohol use, NAFLD by ultrasound and other known liver diseases. We only analysed ALT data for participants without any of these conditions to estimate the mean ALT via a random effects mixed‐model and ULN (95th percentile value) via a bootstrap model with 10,000 resamples. Two investigators independently screened and abstracted the data. Results: From 4,995 studies screened, we identified 60 studies from 1 Jan, 2000 onwards that reported ALT values for 335,163 individuals without NAFLD and viral hepatitis. There was significant heterogeneity for all analyses (all I2 > 96%). The overall mean ALT level for this entire cohort was 20 IU/L and ULN was 32 U/L (Table 1). Several studies also provided subgroup data for ALT by sex, obesity and DM status, allowing for subgroup estimates for “normal” ALT in these groups (Table 1). The mean ALTs were higher in male compared to female (22 versus 18 U/L, p=0.003) and in obese compared to non‐ obese persons (22 vs. 18 U/L, p=0.035). The mean ALT values for DM and non‐DM were similar (p=0.15). Correspondingly, the ULN for ALT was about 32 U/L overall, higher in males (37 vs. 31 U/L) and higher in obese compared to non‐obese persons (36 vs. 29 U/L). There was no statistically significant difference between ALT levels of the DM and non‐DM groups, but there were only four studies providing subgroup data for non‐DM participants. Further, we obtained individual patient level data from 3,834 persons from four centres in Asia (HK, Japan and Taiwan) and found an overall mean ALT of 25 U/L and ULN of 42.0 U/L, with similar trends among the subgroups. Conclusion: We found higher mean ALT in obese compared to non‐obese Asians, and that the ULNs of ALT for male and female were 37 and 30 U/L, higher than the ULNs reported by Prati et al almost 2 decades ago (30 and 19 U/L) for a healthy lean Italian blood donor cohort without metabolic (glucose, lipid) derangement. Given the obesity epidemic and associated metabolic diseases, the ULNs for ALT in the absence of identifiable liver disease including NAFLD should be assessed, though further studies are needed to investigate the long‐term implication of this higher “normal” threshold and to evaluate the same parameters for other racial/ethnic groups.