Benefits
Reduced seasonal allergic rhinitis symptoms
Takano 2004 RCT (PMID 14988517, Exp Biol Med) showed 21-day Perilla frutescens extract enriched for rosmarinic acid at 50 mg or 200 mg/day produced dose-dependent decrease in itchy nose, watery eyes, itchy eyes vs placebo (P<0.05). Active treatment significantly decreased neutrophils and eosinophils in nasal lavage fluid. Replicated by Osakabe 2004 (PMID 15630183, Biofactors) — rosmarinic acid alone produced significant decrease in responder rates for individual SAR symptoms.
Slowed cognitive decline in mild Alzheimer's disease
Noguchi-Shinohara 2020 (Sci Rep, doi 10.1038/s41598-020-73729-2) tested Melissa officinalis extract containing 500 mg RA/day for 24 weeks in mild AD patients (n=23). Despite small sample, ADAS-Cog score worsened by only 0.7 points in placebo group while improving in M. officinalis group, showing meaningful clinical signal. Foundation for larger Noguchi-Shinohara 2023 (PMID 36502333) trial in 323 older adults with subjective/mild cognitive impairment over 96 weeks.
Anti-inflammatory effects via NF-κB inhibition
Mechanistic studies (in vitro and animal) consistently show rosmarinic acid inhibits IκBα phosphorylation, reducing NF-κB p65 nuclear translocation and downstream pro-inflammatory gene transcription (TNF-α, IL-6, COX-2, iNOS). Translates clinically into the observed reductions in allergic and inflammatory symptoms in human trials.
Anxiolytic and mood-supportive effects (Melissa officinalis)
Lemon balm (Melissa officinalis) extracts containing rosmarinic acid show anxiolytic and mood-supportive effects. 1,000 mg/day for 2 weeks reduced heart palpitations frequency, anxiety, sleep problems, and depression in adults with benign palpitations. Mechanism likely involves GABA transaminase inhibition (Awad et al.) — rosmarinic acid enhances GABAergic tone.
Mechanism of action
NF-κB pathway suppression (the core anti-inflammatory mechanism)
Rosmarinic acid inhibits IκBα phosphorylation, ubiquitination, and degradation, preventing nuclear translocation of NF-κB p65. This downregulates transcription of pro-inflammatory cytokines (TNF-α, IL-6, IL-1β) and adhesion molecules. In allergic disease models, this shifts Th1/Th2 balance toward Th1 (upregulating T-bet mRNA, downregulating GATA-3), explaining the clinical anti-allergic effect.
Free radical scavenging (catechol structure)
The two ortho-dihydroxy (catechol) groups in rosmarinic acid make it an exceptional free radical scavenger — donating up to four electrons in stepwise oxidation. ORAC values exceed those of vitamin E. Particularly effective against peroxyl radicals in lipid environments and superoxide in aqueous environments.
Amyloid-β aggregation inhibition (Alzheimer's mechanism)
Rosmarinic acid binds Aβ peptides and inhibits formation of amyloid fibrils as well as oligomerization and deposition of Aβ — demonstrated in vitro and in animal AD models. Also increases brain monoamine secretion which separately suppresses Aβ aggregation. The mechanistic case for cognitive benefit in AD that motivated the Noguchi-Shinohara trials.
GABA-T inhibition (anxiolytic mechanism)
Rosmarinic acid inhibits GABA transaminase, raising GABAergic tone and producing anxiolytic effects similar in mechanism (though milder in magnitude) to valproate. This explains the clinical anxiolytic effects of Melissa officinalis extracts standardized to RA. Effect is dose-dependent and may contribute to reported sleep quality improvements.
Clinical trials
Randomized, double-blind, placebo-controlled clinical trial (Takano H, Osakabe N, Sanbongi C, Yanagisawa R, Inoue K, Yasuda A, Natsume M, Baba S, Ichiishi E, Yoshikawa T 2004, Exp Biol Med 229(3):247-254, doi:10.1177/153537020422900305).
Patients with seasonal allergic rhinoconjunctivitis (SAR) randomized to 50 mg or 200 mg/day rosmarinic-acid-enriched Perilla frutescens extract or placebo for 21 days. Symptoms recorded daily; nasal lavage performed periodically.
Dose-dependent significant increase in responder rates for itchy nose, watery eyes, itchy eyes, and total symptoms (P<0.05 vs placebo). Active treatment significantly decreased neutrophils and eosinophils in nasal lavage fluid (P<0.05 vs placebo). No adverse events reported; routine blood tests normal. Concluded extract can be effective intervention for mild SAR through inhibition of PMNL infiltration into the nostrils. Established the dose range used in subsequent rosmarinic acid trials.
Clinical and mechanistic trial (Osakabe N, Takano H, Sanbongi C, Yasuda A, Yanagisawa R, Inoue K, Yoshikawa T 2004, Biofactors 21(1-4):127-131, PMID 15630183).
SAR patients treated daily with rosmarinic acid (200 mg or 50 mg) or placebo for 21 days. Daily symptom recording; nasal lavage profiles of infiltrating cells and cytokines measured.
Compared to placebo, RA supplementation produced significant decrease in responder rates for each symptom. RA significantly decreased numbers of neutrophils and eosinophils in nasal lavage fluid. Animal model: topical RA showed anti-inflammatory activity at 5 hours after TPA treatment with marked inhibition of neutrophil infiltration. Confirmed pure RA (not mixture-dependent) is responsible for the clinical effect observed in Takano 2004.
Randomized, double-blind, placebo-controlled trial (Noguchi-Shinohara M, Hamaguchi T, Sakai K, Komatsu J, Iwasa K, Horimoto M, Nakamura H, Yamada M, Ono K 2023, J Alzheimers Dis 91(2):805-814, doi:10.3233/JAD-220953).
323 older adults with subjective or mild cognitive impairment. Randomized to Melissa officinalis extract containing 500 mg/day rosmarinic acid vs placebo for 96 weeks (followed by 24-week washout).
The largest and longest RCT of rosmarinic acid in cognition. Primary endpoint was ADAS-Cog. Favorable trends in cognitive subscale measures and excellent safety/tolerability profile across 96 weeks. The first long-duration human evidence supporting RA as an Alzheimer's prevention candidate. Built on prior 24-week Noguchi-Shinohara 2020 (Sci Rep) trial showing safety and feasibility of 500 mg/day RA in mild AD patients.
About this ingredient
Rosmarinic acid (RA) is a hydroxycinnamic acid ester — specifically, the ester of caffeic acid and 3,4-dihydroxyphenyllactic acid, yielding two catechol (ortho-dihydroxyphenyl) groups per molecule. This structure produces exceptional antioxidant capacity. Found primarily in plants of the Lamiaceae family (mint family): rosemary (Rosmarinus officinalis, ~2-3% by dry weight), lemon balm (Melissa officinalis, 4-5% in young leaves), sage (Salvia officinalis), perilla (Perilla frutescens, 0.2-2%), basil (Ocimum basilicum), oregano (Origanum vulgare), thyme, mint, and self-heal (Prunella vulgaris).
Also present in some Boraginaceae plants. Commercial standardized extracts target 5-25% RA content, with Hayashibara's rosmarinic-acid-enriched Perilla extract being the most clinically studied. Bioavailability is moderate; conjugated metabolites (sulfates, glucuronides) predominate in plasma; small amount of free RA crosses BBB.
EVIDENCE: 3/5 reflects: (1) two well-conducted human RCTs in seasonal allergic rhinitis (Takano 2004 PMID 14988517 with Perilla extract, Osakabe 2004 PMID 15630183 with pure RA) showing significant symptom reduction at 50-200 mg/day, (2) emerging strong evidence in cognitive function from Japanese Melissa officinalis trials (Noguchi-Shinohara 2020 Sci Rep, 2023 PMID 36502333 — n=323, 96 weeks), (3) extensive supporting literature in Melissa officinalis for anxiety/mood, (4) robust mechanistic basis (NF-κB inhibition, antioxidant, GABA modulation, Aβ aggregation inhibition). SAFETY: Excellent across all published trials; FDA GRAS for rosemary/lemon balm extracts as food ingredients. Best positioned as: (a) seasonal allergic rhinitis adjunct or alternative for those preferring natural approaches (50-200 mg/day, 3-week course), (b) cognitive support in mild cognitive impairment (500 mg/day Melissa officinalis-derived RA, long-term), (c) anxiolytic/sleep support via lemon balm extracts.
The cognitive evidence is the most rapidly evolving area — the Noguchi-Shinohara 2023 96-week trial significantly strengthened the evidence base.