Eight eligible studies were identified. One eligible study was subsequently excluded; the authors stated that it primarily examined the process and did not assess long-term efficacy.
Seven studies (209 completers) were included in the meta-analysis.
Four studies reported some blinding of the outcome assessor. Six studies were considered to have treatment fidelity. Two studies reported assessor reliability. The authors noted that some studies had high drop-out rates.
There were no statistically significant differences between EMDR and CBT in PTSD symptoms either at post-treatment (ES 0.28, 95% confidence interval, CI: -0.06, 0.63; based on 209 patients) or follow-up (ES 0.13, 95% CI: -0.28, 0.55; based on 163 patients). Statistically significant heterogeneity was detected for both analyses (p=0.03 and p=0.04, respectively).
EMDR was associated with a significant improvement in depression compared with CBT at post-treatment (ES 0.40, 95% CI: 0.05, 0.76; based on 209 patients), but there was no significant difference between treatments at follow-up (ES 0.12, 95% CI: -0.24, 0.48; based on 193 patients). No sstatistically significant heterogeneity was detected for either analysis (p=0.17 and p=0.33, respectively).