Zürrer · Translational psychiatry 2025 · systematic review and meta-analysis of preclinical animal studies · n=82 studies

The translational potential of salvinorin A: systematic review and meta-analysis of preclinical studies.

Cited 5 times in the scientific literature.

Level 5 - mechanism / opinion, no new human data

Systematic review and meta-analysis of preclinical animal research without human data

PubMed 41073402 · doi:10.1038/s41398-025-03638-3 · record verified 2026-08-30

What was done

Authors conducted a systematic review and meta-analysis of preclinical animal studies investigating salvinorin A across Medline, Web of Science, and EMBASE up to June 28, 2024. The review evaluated therapeutic efficacy in neurological and psychiatric disease models, adverse behavioral and physiological effects, and pharmacokinetic parameters.

What was found

From 1,718 identified publications, 82 were included in the qualitative synthesis and 10 in the meta-analysis. In animal models, salvinorin A demonstrated anti-nociceptive, anti-inflammatory, neuroprotective, and anti-addictive effects at doses ranging from 0.1 to 10 mg/kg. Depression model findings were inconsistent, showing both antidepressant and depressogenic outcomes. Adverse effects included anxiogenesis, motor impairment, and cognitive impairment, with minimal vital parameter impact. Pharmacokinetic data showed rapid onset, fast peak, and a half-life of approximately one hour. Sixteen analogues with potentially improved profiles were identified. The abstract reported no quantitative effect sizes.

Why it matters

This review clarifies the preclinical therapeutic landscape of salvinorin A for stroke, pain, and addiction while highlighting that adverse cognitive effects, anxiogenesis, and short half-life represent substantial translational hurdles.

Limits

All included data come from preclinical animal models, preventing direct extrapolation to humans. Only 10 of 82 included studies were suitable for meta-analysis, and specific numerical effect estimates, confidence intervals, and risk of bias metrics are omitted from the abstract.

Cited by