This meta-analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (see additional file 3).
2.1 Literature search strategy
Two authors screened for clinical studies in PubMed (Medline) up to August 2019. Searching terms (medical subject headings [MeSH] terms) was: (“Vascular cell adhesion molecule-1” [MeSH] OR “Intracellular cell adhesion molecule-1” [MeSH] OR “E-Selectin” [MeSH]) AND “preeclampsia” [MeSH]. All articles were retrieved into an EndNote library.
2.2. Inclusion and exclusion criteria
Only studies assessing the relationship between soluble VCAM-1, ICAM-1 or E-selectin and PE were included. Meeting the following four criteria can be included in the meta-analysis: (1) Original prospective, retrospective, cross-sectional or prospective case controlled design are accepted for soluble adhesive molecule concentration detection studies. (2) PE was diagnosed as hypertension (140/90 mmHg or blood pressure elevation 15 mmHg) with proteinuria (300 mg/24 h or 1+ dipstick 6h apart) without pre-existing hypertension. Women with systolic blood pressure >160 mmHg and/or diastolic blood pressure >110 mmHg, accompanied by proteinuria >5 g in a 24-hr collection or with ≥3+ on a dipstick, were defined as severe preeclampsia. Mild pre-eclampsia is defined as fulfilling the criteria for pre-eclampsia but not fulfilling the criteria for severe pre-eclampsia. Controls were healthy pregnant women. (3) Maternal peripheral blood was used for concentration detection. (4)Original data should be included in the study and be represented by mean (SD or SE).
Besides, due to the lack of comparability, duplicate publications, review articles, non-English articles, studies in non-pregnant period, studies on the detection of cytokine concentrations in placenta, peripheral blood cytokines and amniotic fluid, or detecting cytokine concentrations after exogenous test stimulation, or studies provided with case-only were excluded.
2.3 Data extraction
The following information was extracted from each study: author, publication year, study design, and other essential characteristics. For continuity variables, information was collected including the number of case groups and control groups, the mean concentration and standard difference (SD) of maternal plasma or serum VCAM-1, ICAM-1 and E-selectin. If the data provided by the study is standard error (SE), the SE shall be converted into SD according to the formula (SD=SE* ). In addition, ng/ml is the concentration unit of VCAM-1, ICAM-1 and E-selectin for analysis.
2.4 Study quality assessment
Two reviewers independently reviewed the titles/abstracts of all possible studies. Since most articles do not show relevant data in their respective abstracts, the most are reviewed in full. In the event of disagreement, the decision is made by the third reviewer.
2.5 Statistical processing
For the continuous outcomes, we used an inverse-variance weighted mode(18) to calculate the standardized mean difference (MD) and 95% confidence intervals (CIs). Funnel plot(19) and egger's test (20) was used to evaluate publication bias. If there was no heterogeneity (P ≥ 0.1, I2 ≤ 50%), we used the fixed-effects model. If there was a significant heterogeneity (P < 0.1, I2 > 50%), the random-effects model was adopted to allow for the inherent heterogeneity found between studies. Forest plots of the parametric data were generated using the Stata 15.0 software (Stata Corporation, College Station, TX, USA).
We also performed a subgroup analysis of preeclampsia patients. Subgroup 1, mild preeclampsia (M-PE); Subgroup 2, severe preeclampsia(S-PE).