Return
Brain stimulation and VEGF concentrations: meta-analysis, meta-regression, and hierarchical cluster analysis of 200 correlations
A
R
M
M
M
L
D
G
DOI:10.1007/s00406-026-02333-8.png)
Abstract
En 中文
Studies on VEGF responses to brain stimulation (BS) are inconsistent. We systematically reviewed and meta-analyzed these data. Overall meta-analyses for the three main models (time, group, time×group) showed no significant effects (p>0.05). Given the high heterogeneity across interventions and populations, we conducted extensive exploratory subgroup analyses to generate hypotheses. These revealed several within-subgroup effects, but they were based on limited data constraining their interpretability, with several single studies and others on 2–4 observations. For time-difference analyses, higher post-BS VEGF emerged in older rTMS patients (k=4, g=2.14), responders (k=3, g=0.58), and remitters (k=2, g=0.56), but not in other subgroups. Group-difference analyses showed higher post-BS VEGF in tDCS (k=1, g=1.01) and DBS (k=1, g=6.03) groups vs. controls; at baseline, non-responders had lower VEGF than controls (k=3, g=-0.31). Time×group analyses favored rTMS over non-rTMS (k=3, g=1.95), responders over non-responders (k=4, g=0.54), and remitters over non-remitters (k=3, g=0.72). Meta-regression (k=19) identified age as a major variability source (β=1.12, R²=88.9%). HCA ranked the top eight VEGF correlates: PANSS, MADRS, HDRS/HAM-D, HAMA, treatment outcome, age, SHAPS, and BPRS-5. These significant subgroup findings derive from post-hoc exploratory analyses with small sample sizes and should not be interpreted as confirmatory. We provide an early meta-analytical framework, but the small number of observations per BS technique, psychiatric condition, and timepoint precludes definitive conclusions. Future updates with larger, more homogenous datasets are essential to rigorously test whether VEGF can serve as a reliable biomarker of differential BS responses. These exploratory results are hypothesis-generating and require confirmation in future studies.
Keywords:
Biomarker
Brain modeling
Outcome prediction
Precision brain stimulation
Response variability
Vascular age
Journal
IF:
3.7
Papers:
3.4K
Citations:
6.2K
