Objective:  To synthesize meta-analyses of randomized controlled trials on the efficacy of EMDR in reducing anxiety symptoms.

Methods:  Web of Science, PubMed, PsycINFO, and Scopus were searched from inception to April 2026 (PROSPERO CRD420261373395). Two reviewers independently selected studies, extracted data, and appraised quality with AMSTAR 2. Harmonized effect sizes were pooled using random-effects and multilevel second-order meta-analysis, and heterogeneity, moderators, small-study effects, and sensitivity analyses were examined.

Result: Eleven meta-analyses contributing 17 effect sizes were included; one was rated as high quality, one as moderate, three as low, and six as critically low. EMDR was associated with a moderate reduction in anxiety symptoms (g = −0.63, 95% CI [−0.82, −0.45]). Heterogeneity was substantial (I² = 74%) and largely explained by comparator type: effects were large against passive controls (g ≈ −1.01) but small against active psychological treatments (g = −0.36, 95% CI [−0.53, −0.18]). The 95% prediction interval reached approximately null effects. Most comparisons derived from PTSD samples. Results were consistent across sensitivity analyses, including analyses addressing overlap between meta-analyses.

Conclusions: The available evidence suggests that EMDR may reduce anxiety symptoms on average, but its advantage over active treatments appears small. Given the low methodological quality of most included meta-analyses, these findings are indirect and preliminary, and high-quality comparative trials in non-PTSD anxiety disorders are needed.