Search strategy
This study was performed according to the recommendations of the Moose [18]. Electronic databases include PubMed, EMBASE, Ovid, and the Cochrane Library was searched to obtain all the appropriated studies from January 1, 2000 to September 30, 2019. The search keywords used in this search were as follows: NAFLD, Non-alcoholic fatty liver disease, NASH, non-alcoholic steatohepatitis, Normal Alanine Aminotransferase, and Normal ALT. We also manually searched the references of the selected articles to identify additional studies. Only English articles were included in this study.
Inclusion and exclusion criteria
The initially retrieved publications were reviewed by two investigators (Xuefeng Ma and Shousheng Liu) independently. The discrepancy was resolved by discussion with all investigators. Studies that met the following criteria were included: 1) had a case-control, cohort study or cross-sectional analysis; 2) diagnosis the NAFLD by MRI or Ultrasonography and measured the NASH by histology; 3) measured the ALT value on biochemical laboratory; 4) reported the number of NAFLD or NASH patients with normal and abnormal ALT. Studies that met the following criteria were excluded: abstracts, reviews, case reports, and letters. Studies that absence of measurement of the number of NAFLD or NASH patients with normal and abnormal ALT values were also excluded from this study.
Quality assessment and data extraction
The quality of all the included studies was assessed independently by two authors (Xuefeng Ma and Shousheng Liu) using the Newcastle-Ottawa Scale (NOS) tool for case-control and cohort study [19]. We assigned NOS scores of 1-3, 4-6, and 7-9 for low-, intermediate-, and high-quality studies. Cross-sectional analysis was not accessed. The discrepancy was resolved by discussion with all the investigators. The following information was extracted from each study: first author, publication time, the sample size, country, the number of NAFLD or NASH patients with normal ALT, the number of total NAFLD or NASH patents. The data were collected independently by two investigators (Xuefeng Ma and Shousheng Liu).
Statistical analysis and data synthesis
Random-effects model was used to estimate the pooled proportion of NAFLD patients with normal ALT value. Statistical heterogeneity among studies was assessed by Q and I2 statistics. For the Q statistic, heterogeneity was considered to be present when P < 0.1 or I 2> 50%. Publication bias was evaluated visually by funnel plots and the publication bias was considered significant when P value was less than 0.05 in either Begg’s test. The subgroup was carried out by region, type of study, outcome measurement and group size. We used the metaphor (version 2.0-0) and meta (version 4.9-5) packages of R (version 3.6.1) to conduct the different analyses [20] and all statistical tests.
Results
Study characteristics
The flow diagram of studies selection was shown in the Figure 1. A total of 699 references were identified according to our search strategy. After removed the duplication, reviews, animal studies, and irrelevant resources, 34 potential studies were selected to further evaluation. After excluded 23 improper studies, a total of 11 studies with 4084 patients [13-15, 21-28] which matched the inclusion criteria were included in this meta-analysis.
The main features of the included studies were shown in Tables 1, which including 6 retrospective cohort studies, 3 prospective cohort studies and 2 cross-sectional analyses. Among these studies, subjects in 6 studies were NASH patients. Among all of the included studies, 4 studies are from North America, 3 from Europe and 4 from Asia. NOS scores suggested that 9 studies possessed the high quality with all the NOS scores were 7. We did not assess the quality of cross-sectional studies.
Proportion of NAFLD patients with normal ALT in overall NAFLD patients
To investigate the pooled proportion of NAFLD patients with normal ALT overall NAFDL patients, 11 studies from different countries and regions which include 4 from the USA, 1 from Italy, 1 from India, 1 from Turkey, 1 from the UK, and 3 from China were included. The overall pooled proportion were 0.25 (95%CI: 0.20 - 0.31) according to the calculation by random-effects model (P < 0.001, I2 = 92.0%) (Figure 2).
Subgroup analysis was performed to explore the sources of heterogeneity. We evaluated potential sources of heterogeneity between region, type of study, diagnostic methods and group size (Table 2). The summarized proportion of NAFLD patients with normal ALT value in Asia was 0.30 (95%CI: 0.25 - 0.35, I2 = 52.0%), higher than in North America 0.24 (95%CI: 0.16 - 0.36, I2 = 97.0%) and Europe 0.19 (95%CI: 0.14 - 0.26, I2 = 72.0%). The summarized proportion of NAFLD patients with normal ALT value in prospective cohort study group was 0.32 (95%CI: 0.22 - 0.43, I2 = 85.0%), higher than in retrospective cohort study group 0.21 (95%CI: 0.10 - 0.28, I2 = 87.0%) and cross-sectional analysis group 0.27 (95%CI: 0.19 - 0.37, I2 = 85.0%). The summarized proportion of NAFLD patients with normal ALT value in MRI diagnostic group was 0.43 (95%CI: 0.39 - 0.38), higher than in the Histology diagnostic group 0.22 (95%CI: 0.17 - 0.28, I2 = 85.0%) and Ultrasonography diagnostic group 0.26 (95%CI: 0.20 - 0.34, I2 = 83.0%). The summarized proportion of NAFLD patients with normal ALT value in more than 300 size group was 0.27 (95%CI: 0.23 - 0.33, I2 = 97.0%), higher than in less than 300 size group 0.22 (95%CI: 0.14 - 0.31, I2 = 46.0%). Funnel plots was constructed to investigate the publication bias, and the results shown that no publication bias was exist (P = 0.14) (Figure 3).
Sensitivity analysis was carried out to evaluate the influence of a single study on the results of this meta-analysis. We found that no significant changed was observed of I2 values when any one study was removed from this meta-analysis.
Summarized proportion of NASH patients with normal ALT in overall NASH patients
6 of the 11 studies with 1023 patients were selected for the meta-analysis of the summarized proportion of NASH patients with normal ALT in overall NASH patients [23-28]. As shown in the results, the summarized proportion NASH patients with normal ALT in overall NASH patients was 19% (95%CI: 13% - 27%), calculated with the random-effects model (P < 0.001, I2 = 85.0%) (Figure 4).
The subgroup analysis was used to explore the sources of heterogeneity. We evaluated the possible sources of heterogeneity between studies, including region, type of study, diagnosis method and group size (Table 3). The summarized proportion of NASH patients with normal ALT value in North America was 0.25 (95%CI: 0.19 - 0.33), higher than in Asia 0.04 (95%CI: 0.01 - 0.24, I2 = 85.0%) and Europe 0.19 (95%CI: 0.11 - 0.30, I2 = 81.0%). The summarized proportion of NASH patients with normal ALT value in retrospective cohort study group 0.21 (95%CI: 015- 0.29, I2 = 84.0%), higher than in prospective cohort study group was 0.04 (95%CI: 0.00 - 0.21). The summarized proportion of NASH patients with normal ALT value in MRI diagnostic group was 0.30 (95%CI: 0.25 - 0.37), higher than in histology diagnostic group 0.17 (95%CI: 0.12 - 0.25, I2 = 77.0%). The summarized proportion of NASH patients with normal ALT value in more than 300 size group was 0.20 (95%CI: 0.12 - 0.32, I2 = 91.0%), equal to the proportion in less than 300 size group 0.20 (95%CI: 0.15 - 0.26, I2 = 0%).