We conducted a comprehensive literature search to identify studies that examined VEGF levels in the circulation (serum or plasma) of patients with BD compared to healthy controls, as well as studies exploring the association between VEGF gene polymorphisms and BD. The search covered the MEDLINE, Embase, and Cochrane databases and included articles from the inception of these databases up to July 2024. Our search strategy involved keywords such as “vascular endothelial growth factor” “serum OR plasma OR circulating,” “polymorphism,” and “Behçet’s disease.” Additionally, we reviewed the references in the identified articles to find any relevant studies that may not have been captured in the database search. Eligible studies met the following criteria: (1) case-control, cohort, or cross-sectional design; (2) provided data on circulating VEGF levels in both BD patients and control groups; and (3) examined VEGF gene polymorphisms in both BD and control groups. Studies were excluded if they had overlapping or insufficient data, or if they were review articles or case reports. Two independent reviewers extracted information on the methods and results from the original studies, resolving any discrepancies by consensus. Extracted data included details such as the primary author, year of publication, country, ethnicity, adjustments for age and sex, number of participants, mean and standard deviation (SD) of VEGF levels, and allele and genotype frequencies of VEGF gene polymorphisms. Data were reported as mean and SD, or as median and interquartile range, using established formulae [20,21]. The definitions of active versus inactive BD varied across the studies: Arica et al. [6]: Patients with at least one mucocutaneous lesion or active organ involvement were classified as having “active disease,” while those without any mucocutaneous lesions or systemic complaints at the time of blood collection were defined as having “inactive disease.” Gheita et al. [7]: Active disease was defined as a Behçet’s Disease Current Activity Form score of 2 or higher. Yalçındağ et al. [11]: Ocular involvement was considered an indication of active disease. Ibrahim et al. [12]: Patients with at least two of the following were classified as having active disease: aphthous stomatitis, genital ulceration, positive pathergy test, skin lesions, anterior iridocyclitis, panuveitis or posterior vasculitis, arthritis, vascular lesions, or manifestations affecting the pulmonary or central nervous systems. Kamoun et al. [14]: Ocular involvement was considered an indicator of active disease. Shaker et al. [15]: Active disease was defined by the presence of three out of five major findings on admission (skin lesions, positive pathergy test, uveitis, oral ulcers, and genital ulcers). Patients without any clinical or laboratory signs of BD for at least 1 month were classified as having inactive disease. Cekmen et al. [17]: The presence of three out of five major findings on admission (skin lesions, positive pathergy test, uveitis, oral aphthae, and genital ulcers) was considered to indicate active disease. The quality of each study was assessed using the Newcastle-Ottawa Scale [22]. The meta-analysis was conducted following the PRISMA guidelines [23].

We performed a meta-analysis to examine the relationship between VEGF levels and BD, as well as to assess the impact of minor alleles versus major alleles in various VEGF gene polymorphisms. Results were expressed as standardized mean differences (SMDs) with 95% confidence intervals (CIs) for continuous data, and odds ratios (ORs) with 95% CIs for dichotomous data. Cochran’s Q test was used to assess within-study and between-study variations and to evaluate heterogeneity. A heterogeneity test determined whether the studies were evaluating the same effect. If the Q-statistic was significant (p<0.10), indicating heterogeneity, a random-effects model was applied; otherwise, a fixed-effects model was used, assuming all studies estimated the same underlying effect and considering within-study variations [24]. Heterogeneity was quantified using I², which measures inconsistency between studies, with I² values indicating low, moderate, or high levels of variation [25]. Statistical analyses were conducted using the Comprehensive Meta-Analysis Program (BioStat, Englewood, NJ, USA).

To identify potential sources of heterogeneity observed in the meta-analysis, we conducted a meta-regression analysis considering variables such as ethnicity, adjustments for age and/or sex, publication year, sample size, and data type. Sensitivity analysis was performed by systematically omitting each study to assess its influence on the overall effect size. Unlike funnel plots, which can be subjective and require studies of various types and sample sizes, Egger’s linear regression test was used to assess publication bias, detecting funnel plot asymmetry by using a natural logarithm of the effect size [26].

We identified a total of 128 studies through electronic and manual searches. Out of these, 19 were selected for a thorough review based on their titles and abstracts. Four studies were excluded due to either insufficient data or duplicate information. Consequently, 15 studies met the inclusion criteria which included data from 885 patients with BD and 875 control subjects (Tables 1, 2, and Figure 1) [5-19]. Two of these studies provided data on both VEGF levels and the VEGF gene polymorphisms of interest [5,14] and another two studies provided only data on the VEGF gene polymorphisms [18,19]. Thus, 13 investigated VEGF levels in both the affected and control groups, while four focused on VEGF gene polymorphisms in both BD and control groups (Tables 1 and 2). Meta-analyses were conducted on VEGF polymorphisms where there were at least two comparisons. Due to the limited number of studies on candidate gene associations, three separate meta-analyses were performed, specifically analyzing VEGF polymorphisms –634 C/G, +936 C/T, and 18 bp insertion/deletion (I/D) at –2549. Each study was rated for quality on a scale of 5 to 7. Details regarding the characteristics of the included studies are provided in Tables 1 and 2.

The meta-analysis showed a significant difference in circulating VEGF levels between BD patients and controls (SMD=1.726, 95% CI=1.030~2.421, p<0.001) (Table 3 and Figure 2). Stratified by ethnicity, BD patients of European and Arab descent had higher VEGF levels compared to controls (Table 3). Subgroup analysis confirmed elevated VEGF levels in BD patients regardless of data type and sample size (Table 3). Furthermore, VEGF levels were significantly higher in patients with active BD compared to those with inactive BD (SMD=0.635, 95% CI=0.092~1.177, p=0.022) (Table 3 and Figure 2).

The meta-analysis revealed no significant association between BD and the VEGF –634 C allele when comparing pooled data from affected individuals and controls (OR=1.023, 95% CI=0.707~1.481, p=0.904) (Table 4 and Figure 3). Similarly, no significant association was found between BD and the VEGF +936 T allele (OR=1.072, 95% CI=0.811~1.417, p=0.628) (Table 4 and Figure 3). The analysis showed no significant association between BD and the VEGF 18 bp I/D polymorphism at –2549 (Table 4 and Figure 3).

Heterogeneity was evident in most meta-analyses of VEGF levels in BD patients (Table 3). Meta-regression analysis indicated that data type, sample size, and publication year contributed to this heterogeneity. Sensitivity analysis showed that no single study had a disproportionate effect on the pooled SMD, supporting the robustness of the meta-analysis results. Heterogeneity was also present in meta-analyses of VEGF polymorphisms (Table 4). Although publication bias can arise from an imbalance of positive findings, our analysis found no evidence of such bias. Funnel plots did not show asymmetry, and Egger’s regression test yielded p-values >0.05.