2.1 Protocol:
We conducted this meta-analysis strictly according to the MOOSE list, and the contents of the system review and meta-analysis have been well reported in the complete MOOSE list (S1 Table).
2.2 Search strategy:
Two investigators (HD,SZD) used advanced “glycation end products” OR “AGEs” OR “advanced glycosylation end products” OR “glycosylation end products, advanced” [MeSH] OR "glycation end products receptor" OR "RAGE" AND "obesity"[MeSH] OR "overweight" as the search strategy. Since we retrieve the following database from the database to the data on May 20, 2021: PubMed, EMBASE, Web of Science, the Cochrane Library. Furthermore, references of identified citations and Google Scholar were also searched to identify additional studies. Truncation was used when appropriate to fine-tune the search and increase the number of relevant findings.
2.3 Selection of articles:
The retrieval data of different expression formats were imported into Endnote version X9 software (Thomson Reuters, Stamford, CT, USA) for integration, and the same research from different database sources was initially deleted through the "find duplicates" function. The study will be included according to the following criteria: RAGE was detected by blood serum in study; the study should be case-control study; the article was in English in a peer-reviewed journal. The study will be included according to the following criteria: The study designed was animal experiments、reviews, conference articles and the study was not divided into obesity group and control group. Two authors (DH, SZD) excluded studies obviously unrelated to the purpose of the article by reading the title and abstract, which did not study “obesity” and “RAGE”. Then screening eligible study independently selected for full-text review. Divergences were solved by the consensus that was reached in advance. If >1 article was published on the same clinical trial, we chose the article with the most complete data. When the research data is incomplete, we try to contact the author of the article to acquire the whole data.
2.4 Data extraction:
Two researchers (DH, SZD) independently extracted the relevant data of each study, including: the first author of the study, the year of publication, Country of publication, Publication type, the number of participants, age, body mass index and gender in obesity group and non-obesity group were compared, diagnostic criteria of obesity, matching principle, sRAGE value (Table 1). The form of conversion of RAGE is different in these articles. We convert units to make them consistent by using this formula “1 ng/ml=1000 pg/ml”. Some of the data are expressed by Inter-quartile range, while others are showed by the sample mean and standard deviation. In order to keep the consistency of data, we used the calculation method of Professor Tong Tiejun to estimate the sample mean and standard deviation from the sample size, median, range and/or interquartile range (http://www.math.hkbu.edu.hk/~tongt/papers/median2mean.html).
2.5 Literature quality evaluation:
The Newcastle-Ottawa Scale (NOS)15 will be used to evaluate biases in studies included in this meta-analysis. This scale is evaluated from three aspects, including object selection, comparability, outcome, and exposure. By comparing this scale, all studies will be divided into three quality grades based on score: high, and low, and high-quality articles score more than 5 stars, and the lower is 0-5 stars.
2.6 Data analysis:
The main outcome of our study is the relationship between serum advanced glycation end product receptor and obesity. According to the characteristics of the subjects, we will conduct subgroup analysis based on age, study area, obesity degree, obesity type, complications, etc.
Rev-man 5.3 software (Cochrane organization, England) will be used for data analysis of the study. In this meta-analysis, we used means and SDs to measure the continuous results of changes in sRAGE. The random effects model will be used for combination analysis of data and reported as mean difference (MD) and 95% confidence interval. I2 statistics is used to evaluate the heterogeneity of meta-analysis, and when it is more than 50%, there may be substantial heterogeneity. "leave one out" is used for sensitivity analysis to evaluate the robustness of the overall results. When more than 10 studies were included in the meta-analysis, publication bias was necessary. Funnel visualization analysis and egger's regression test were used to test publication bias. In all analyses, P values (two sides) < 0.05 showed statistical significance.