Boswellia

Best-in-class: Boswellia Phytosome

Frankincense extract standardized to boswellic acids, in an absorption-enhanced phytosome, targeting the 5-LOX inflammatory pathway (distinct from NSAIDs' COX pathway).

How Boswellia actually works

Boswellic acids, principally AKBA, inhibit 5-lipoxygenase — a completely different arm of the eicosanoid cascade from the COX enzymes that NSAIDs and curcumin target. That is why it stacks with curcumin rather than duplicating it: two branches of the same inflammatory tree. The phytosome formulation addresses boswellic acids' poor absorption.

Where to get Boswellia

The evidence for Boswellia

Graded by what exists behind each claim.

✅ Clinically validated

• RCTs show reduced osteoarthritis pain and improved joint function.
• Signals for benefit in inflammatory bowel and airway conditions.

📊 Correlative data

• Traditional anti-inflammatory use aligns with the modern joint-pain trials.

🧪 Theoretical / extrapolated benefits

• 5-LOX inhibition is proposed to complement COX-targeting anti-inflammatories — mechanistically sound, broader claims less proven.

How to read these tiers: they say how much human evidence exists, not how well something works — and ✗ flags harm, never a disappointing trial. How the evidence tiers work →

What Boswellia actually does

Boswellic acids are pentacyclic triterpene acids from the oleogum resin of Boswellia serrata, and there are six of them worth naming because a quality survey measures exactly these: KBA, AKBA, alpha-BA, beta-BA, acetyl-alpha-BA and acetyl-beta-BA Meins 2016. The one the marketing is about is AKBA — 3-acetyl-11-keto-beta-boswellic acid.

The founding experiment gave a number and a specificity claim in the same paper. Among the boswellic acids, AKBA produced the most pronounced inhibition of 5-lipoxygenase product formation, with an IC₅₀ of 1.5 micromolar — and boswellic acids tested at concentrations up to 400 micromolar did not impair cyclooxygenase or 12-lipoxygenase Safayhi 1992. A 260-fold separation between the target and the two nearest enzymes is what ‘specific’ means quantitatively, and the same work established the inhibition was non-redox: not an antioxidant quenching the enzyme's iron chemistry, but a molecule binding a site.

The structure–activity work then showed which parts of the molecule the site cares about. Removing the acetate raised the IC₅₀ from 1.5 to 3 micromolar in intact cells and from 8 to 20 micromolar cell-free; reducing the carboxyl group to an alcohol raised it to 4.5 and 45 micromolar respectively Sailer 1996. Two functional groups, each worth a two- to five-fold change — that is a defined binding pocket, not a general triterpene effect. Note also that AKBA is more potent in intact cells (1.5 microM) than against the isolated enzyme (8 microM), which means the lipophilic acid concentrates inside cells rather than being excluded from them.

Then a protein in the blood takes the whole mechanism away. When 11-keto-boswellic acids were re-examined, they suppressed 5-LOX product formation in isolated human neutrophils with IC₅₀ values of 2.8 to 8.8 micromolar — and over 95% of them bound to albumin at physiological concentrations, which abolished the inhibitory effect Siemoneit 2009. That is the killer fact about this molecule and it is a completely different failure mode from a molecule that never gets absorbed: boswellic acid arrives in plasma perfectly well and is then held by a carrier protein at a free concentration far below the one that inhibits the enzyme.

And there is a second molecular target that the card explicitly denies. Boswellic acids were identified as direct inhibitors of cyclooxygenase-1, with an IC₅₀ of 6 micromolar in intact human platelets and 32 micromolar against the isolated enzyme, with COX-2 inhibited less efficiently Siemoneit 2008. Read that against the card, which sells boswellia as targeting a pathway “distinct from NSAIDs' COX pathway” and therefore safe to stack. In intact human platelets, boswellic acids inhibit COX-1 at a lower concentration than 11-keto-boswellic acids inhibit 5-LOX in neutrophils. COX-1 in platelets is the enzyme aspirin acetylates. That single number converts the vague ‘mild antiplatelet activity’ caution into a named enzyme with a potency, and it undermines the stacking argument the product is sold on.

Cell, rodent, human — and where it stops

Enzyme and cell: strong, human, and quantified. AKBA at IC₅₀ 1.5 microM against 5-LOX product formation with COX and 12-LOX untouched to 400 microM Safayhi 1992; a defined structure–activity series Sailer 1996; COX-1 at 6 microM in intact human platelets Siemoneit 2008. Every one of those is human material.

Then the human pharmacodynamic test, and it failed. Healthy volunteers took a single 800 mg oral dose of frankincense extract, and it failed to suppress plasma leukotriene B₄ Siemoneit 2009. LTB4 is the direct product of the pathway AKBA is supposed to block. The authors' own conclusion was that boswellic acids are direct 5-LOX inhibitors in vitro but that the pharmacological relevance of that interference in vivo seems questionable. That is the mechanism's own literature retiring the mechanism.

And yet the clinical trials are positive, which is the interesting part. Seven randomized trials in 545 patients were pooled. Pain on a visual analog scale improved by a weighted mean difference of −8.33 (95% CI −11.19 to −5.46, P < 0.00001); WOMAC pain by −14.22 (−22.34 to −6.09, P = 0.0006); WOMAC stiffness −10.04; WOMAC function −10.75; the Lequesne index −2.27; and the pooled analysis concluded that at least four weeks of treatment is needed Yu 2020. The individual trials are consistent with that: 5-Loxin at 100 and 250 mg a day produced clinically and statistically significant improvements in pain and physical function, with the 250 mg arm separating from placebo by seven days Sengupta 2008.

So the specific obstacle is the most unusual one on this site: the outcome data survive and the mechanism does not. A compound that does not measurably inhibit its target pathway in vivo Siemoneit 2009 nevertheless reduces osteoarthritis pain across seven randomized trials Yu 2020. There are three honest explanations and no way yet to choose between them. One: the real target is COX-1, which is inhibited at concentrations closer to plausible Siemoneit 2008 — in which case boswellia is a slow NSAID and should be described as one. Two: the relevant target is something else entirely that nobody has assayed. Three: the trials, which are mostly small and industry-adjacent, are measuring a symptom endpoint with a large placebo component. The page that pretends to know which is lying.

The obstacle underneath all three is exposure, and it has a practical fix. Food changes it: boswellic acid bioavailability increased several-fold with a high-fat meal compared with fasted dosing Sterk 2004. That is a free intervention, it is larger than most formulation claims, and almost nobody is told about it.

Boswellia — which form, and does it matter

Start with what this bottle actually is, because the card's dose is wrong. The Boswellia Phytosome label filed with the NIH database lists Casperome — an Indian frankincense phytosome from Boswellia serrata — at 350 mg per capsule, one capsule twice daily, which is 700 mg a day, and it declares no boswellic acid percentage and no AKBA percentage at all Office of Dietary Supplements 2025. The card says 100–250 mg of a bioavailable form daily. The label says seven hundred.

The standardization problem is measured, and it is the worst quality finding in this cohort. The top-selling European and American boswellic acid supplements were assayed. 41% did not comply with their label declaration. One Italian product contained none of the six characteristic boswellic acids. Another US product contained only traces. And two products contained manipulated extract enriched to up to 66% AKBA, which was not declared on the label Meins 2016. Two failure directions at once: capsules with nothing in them, and capsules with an undeclared twenty-fold enrichment of the most active constituent.

Which makes the AKBA percentage the number to buy on, and explains why the trial extracts state it. 5-Loxin is standardized to 30% AKBA and it is the material with the randomized trial Sengupta 2008. Natural gum resin runs at a few percent AKBA. A product declaring ‘65% boswellic acids’ without an AKBA figure has told you about the fraction that includes the less active members Sailer 1996, not about the one the pharmacology belongs to.

The phytosome is a real formulation with a real measurement. The lecithin delivery form of Boswellia extract raised plasma exposure up to 7-fold for KBA and 3-fold for beta-BA by AUC against the non-formulated extract, and produced a 35-fold increase in brain KBA/AKBA concentration and up to 17 times higher boswellic acid levels in poorly vascularized tissue such as the eye Husch 2013. Seven-fold is a genuine pharmacokinetic gain, and the tissue-distribution numbers are larger still.

But hold that against the albumin number before deciding it settles anything. Over 95% of 11-keto-boswellic acid is albumin-bound at physiological concentrations, and that binding abolished 5-LOX inhibition Siemoneit 2009. Raising total plasma concentration seven-fold raises the free fraction seven-fold too, from a very small number — and the requirement is 2.8–8.8 microM of free drug at the enzyme. The phytosome is the best-characterized boswellia formulation on the market and still nobody has published a free, unbound boswellic acid concentration in a dosed human. That is the measurement this whole category is missing.

Practical instructions, then. Buy on a declared AKBA percentage Meins 2016; take it with a fatty meal, which is worth several-fold on its own Sterk 2004; and give it at least four weeks before judging Yu 2020.

What would have to be true, and how you would know it was not

1. womac pain, with the pooled number rather than a promise. Score WOMAC pain before starting and at eight weeks. Predict a fall of about 14 points on the 0–100 normalized scale, with function down about 11 and stiffness about 10 Yu 2020. Those are pooled weighted mean differences against control from seven trials in 545 patients, which makes them the most defensible effect-size predictions on any joint product in this catalog.

2. Timing, because it is diagnostic rather than decorative. Predict the first change at one to two weeks — the 250 mg arm of the 5-Loxin trial separated from placebo by seven days Sengupta 2008 — and the full effect by four weeks Yu 2020. Nothing at all by six weeks on a properly standardized product taken with food is a genuine failure, not a reason to persevere.

3. Bleeding time and platelet function — the prediction that cuts against the product, and it follows from a named enzyme. Boswellic acids inhibit COX-1 in intact human platelets at an IC₅₀ of 6 micromolar Siemoneit 2008. Predict a measurable prolongation of platelet function analyzer closure time, or increased bruising, on 700 mg a day of a phytosome formulation Office of Dietary Supplements 2025 that raises plasma exposure several-fold Husch 2013. If that holds, the card's central selling point — a different pathway from NSAIDs, so they stack safely — is wrong, and boswellia should be stopped before surgery the way an NSAID is.

4. hs-crp, and the honest prediction is nothing. Predict no change at eight weeks. Osteoarthritis is not a high-CRP disease, none of the pooled trials reported a systemic inflammatory marker Yu 2020, and a single 800 mg dose failed to move the one eicosanoid it should have moved Siemoneit 2009. A flat CRP alongside a 14-point WOMAC fall is the expected and interesting result: symptom relief without a systemic inflammatory signature.

5. The prediction that would settle the mechanism, and anyone with access to a lab could run it. Measure plasma leukotriene B₄ or urinary LTE₄ before and two hours after a dose of a phytosome product. Predict no suppression, because 800 mg of a conventional extract produced none Siemoneit 2009. If a modern formulation does suppress it Husch 2013, the 5-LOX mechanism is rescued and this section is wrong — which is exactly why it is the experiment worth doing.

What nobody has tested yet

Nobody has published a free, unbound boswellic acid concentration in a dosed human. Over 95% is albumin-bound and that binding abolishes the enzyme inhibition Siemoneit 2009; every published pharmacokinetic figure, including the phytosome's seven-fold gain, is a total concentration Husch 2013. Equilibrium dialysis on stored plasma from an existing study would produce the number in an afternoon, and it decides whether the marketed mechanism is reachable at any dose.

Nobody has tested the phytosome against the leukotriene endpoint. The negative human pharmacodynamic result used a conventional 800 mg extract Siemoneit 2009; the formulation that raises exposure several-fold has never been put through the same test Husch 2013. One crossover study in twelve volunteers, with plasma LTB4 as the read-out, would tell the field whether formulation solved the problem or merely raised a number.

Nobody has run boswellia head to head against an NSAID with a platelet endpoint. If COX-1 in platelets is inhibited at 6 micromolar Siemoneit 2008, then the correct comparator is not placebo, it is low-dose aspirin, and the correct safety endpoint is platelet aggregation rather than a questionnaire. That trial does not exist, and it is the one that would tell surgeons what to do with this.

And nobody has re-run the quality survey. Forty-one percent of top-selling products failed their own label declaration, and one contained none of the six characteristic boswellic acids Meins 2016. That was a decade ago. Repeating it on today's shelf — including the phytosome products, which declare no boswellic acid content at all Office of Dietary Supplements 2025 — is a week of chromatography and would be the most useful consumer study anyone could publish on this ingredient.

Boswellia — its own safety story, not its category's

The antiplatelet caution on this product is not a class courtesy; it has an enzyme and a number. Boswellic acids inhibit cyclooxygenase-1 in intact human platelets with an IC₅₀ of 6 micromolar Siemoneit 2008. Platelet COX-1 is the aspirin target, and it is the enzyme whose irreversible inhibition is why aspirin is stopped before surgery. Anyone on aspirin, clopidogrel, warfarin or a direct oral anticoagulant is stacking a second COX-1 inhibitor, and anyone facing a procedure should stop boswellia on the same schedule they would stop an NSAID rather than on the vague two-week supplement rule.

Which also corrects the card's stacking advice. The card recommends boswellia specifically because it hits a different pathway from NSAIDs and therefore pairs well with them. The molecular work says COX-1 is a direct target Siemoneit 2008, so the combination is additive on the enzyme most responsible for NSAID gastric injury rather than complementary to it. That does not make boswellia dangerous; it makes the specific reason given for combining it wrong.

The quality risk is this product category's own, and it runs in both directions. Forty-one percent of top sellers did not comply with their label declaration; one contained no characteristic boswellic acids at all; and two contained an undeclared extract enriched to 66% AKBA Meins 2016. An undeclared twentyfold enrichment is the more concerning of the two, because a person who titrated their dose against a weak product and then re-ordered from a different batch has changed their exposure by an order of magnitude without changing the number of capsules.

The formulation raises tissue exposure well beyond plasma, and that deserves a caution nobody gives. The lecithin formulation produced a 35-fold increase in brain KBA/AKBA and up to 17-fold higher levels in poorly vascularized tissue such as the eye Husch 2013. Those are exactly the compartments where conventional safety data do not exist. Nothing in this file reports harm from that, and nothing reports looking for it — a supplement engineered to cross into the central nervous system is a different product from a gum resin, and the safety literature has not caught up with the formulation.

Two smaller ones and one that is properly reassuring. Boswellic acids interact with cytochrome P450 enzymes, so a several-fold increase in systemic exposure from a phytosome is the context in which a drug interaction becomes plausible rather than theoretical. Pregnancy is an avoid, on traditional emmenagogue use and the absence of any data. And the reassuring part: across seven randomized trials in 545 patients the pooled conclusion was that this is an effective and safe option for osteoarthritis Yu 2020, and the 5-Loxin trial reported safety parameters unchanged against placebo Sengupta 2008. The concerns above are about the enzyme it really hits and the bottle you really bought, not about toxicity.

Sources read for this page

• Office of Dietary Supplements, National Institutes of Health. Boswellia Phytosome (Thorne) -- filed Supplement Facts panel: Casperome Indian frankincense phytosome from Boswellia serrata 350 mg per capsule, one capsule twice daily, no boswellic acid or AKBA percentage declared; on market. NIH Dietary Supplement Label Database, label recorded 2025
• Safayhi H, et al. Boswellic acids: novel, specific, nonredox inhibitors of 5-lipoxygenase. Journal of Pharmacology and Experimental Therapeutics 1992;261(3):1143-1146 · PMID 1602379
• Sailer ER, et al. Acetyl-11-keto-beta-boswellic acid (AKBA): structure requirements for binding and 5-lipoxygenase inhibitory activity. British Journal of Pharmacology 1996;117(4):615-618 · PMID 8646405
• Siemoneit U, et al. On the interference of boswellic acids with 5-lipoxygenase: mechanistic studies in vitro and pharmacological relevance. European Journal of Pharmacology 2009;606(1-3):246-254 · PMID 19374837
• Siemoneit U, et al. Identification and functional analysis of cyclooxygenase-1 as a molecular target of boswellic acids. Biochemical Pharmacology 2008;75(2):503-513 · PMID 17945191
• Sterk V, Buchele B, Simmet T. Effect of food intake on the bioavailability of boswellic acids from a herbal preparation in healthy volunteers. Planta Medica 2004;70(12):1155-1160 · PMID 15643550
• Sengupta K, et al. A double blind, randomized, placebo controlled study of the efficacy and safety of 5-Loxin for treatment of osteoarthritis of the knee. Arthritis Research and Therapy 2008;10(4):R85 · PMID 18667054
• Meins J, et al. Survey on the Quality of the Top-Selling European and American Botanical Dietary Supplements Containing Boswellic Acids. Planta Medica 2016;82(6):573-579 · PMID 27054914
• Husch J. Enhanced absorption of boswellic acids by a lecithin delivery form (Phytosome) of Boswellia extract. Fitoterapia 2013 · PMID 23092618
• Yu G, et al. Effectiveness of Boswellia and Boswellia extract for osteoarthritis patients: a systematic review and meta-analysis. BMC Complementary Medicine and Therapies 2020;20(1):225 · PMID 32680575

How you would know if it worked

Boswellia — safety & side effects

Bloodwork to run alongside Boswellia

Baseline first, then again at 8–12 weeks.

The Longevity Baseline panel covers these in one order — 13 markers, $219.10 with the discount applied.

Check results you already have → · All 103 markers A–Z

Boswellia — frequently asked questions

Boswellia inside a finished plan

One arm of 2 Protocol Blueprints, free to read in full.

What Boswellia is used for

Boswellia appears under 3 goals in the goal router.

Related Longevity & Antioxidants supplements

← Browse the full Supplement Vault

↑ Back to on this page