Skip to content
Ejaculation.org

Meta-analysis · 2025

Effect of Yoga in Premature Ejaculation: A Systemic Review and Meta-analysis

Rohilla J, Hasan S, Kumar V, Dhiman V, Vardhan G (2025). Effect of Yoga in Premature Ejaculation: A Systemic Review and Meta-analysis. International Journal of Yoga. doi:10.4103/ijoy.ijoy_52_25

Key finding
This review found only 5 studies of yoga for PE; pooling 3 showed a positive effect (SMD 1.74) but with very high heterogeneity (I² 90.7%), small samples and no standard yoga protocol, so larger randomised trials are needed.

Summary

The authors searched five databases from 1990 to January 2025 for studies of yoga for premature ejaculation (PE). Five studies met the inclusion criteria, and three that measured intravaginal ejaculatory latency time (IELT, the time from penetration to ejaculation) were pooled with a random-effects model.

The pooled analysis showed a significant positive effect of yoga (standardised mean difference 1.74, 95% CI 1.27–2.22) but with very high heterogeneity (I² = 90.7%), meaning the studies' results differed a lot. In the one study comparing yoga with fluoxetine, latency improved more with yoga (79.6 vs 34.2 seconds). No significant side effects were reported in the yoga groups. The authors suggest yoga may work through stress reduction, autonomic nervous system changes and pelvic floor strengthening, and call for larger, well-designed randomised trials.

Limitations

Only a handful of small studies exist, several not randomised, mostly from India and with different yoga programs. Blinding is difficult in yoga trials. High heterogeneity makes the pooled number unreliable as an estimate of the real effect. The review was written by authors who had published one of the yoga studies in this field.

What it means for you

The direction of the evidence is positive, but it rests on very few small trials. Yoga is low-risk and can support relaxation, breathing and pelvic floor awareness; treat it as a complement to tested techniques rather than a proven treatment.