Supplements
boswellia decreases inflammation
In plain terms: Is boswellia actually anti-inflammatory?
Strong support Supplements 💰 Industry COI noted
Part of: 🧪 boswellia
RefutedContestedStrong support
consensus score 1.00
Yes, with a real mechanism — its boswellic acids block a key inflammation enzyme (5-LOX), and a human trial showed it reduced brain swelling in cancer patients. The mechanism is solid; broad human 'lowers inflammation' data are still limited.
📅 Last reviewed: 2026-07-15 ⓘ
Evidence ladder
How far up the ladder this claim has climbed. A high consensus on a low rung means "consistent so far," not "proven in people."
Top evidence so far: Human trials (RCT / n-of-1)
MechanismIn-vitroAnimalObservationalRCTMeta-analysis
How the studies fall
10 support 0 contradict 0 tested null 0 mixed · 10 sources, 10 independent groups
What the evidence shows
Boswellia's textbook mechanism — boswellic acids (esp. AKBA) inhibiting 5-lipoxygenase to cut leukotrienes — is CONTESTED, not settled: the pharmacokinetic review (Abdel-Tawab 2011) reports boswellic acids failed to inhibit leukotriene formation in human whole blood and calls 5-LO inhibition 'questionable' at achievable plasma levels, against Ammon's mechanism reviews.
The evidence (19)
| Source | Grade | Stance | Quality | Finding |
|---|---|---|---|---|
| Ammon 2010 · Phytomedicine | mechanism | supports | low | Narrative review, no new data. Every inflammation-relevant mechanism it reports runs anti-inflammatory: BAs 'inhibit activation of NFkappaB', with 'a down regulation of TNF-alpha and decrease of IL-1, IL-2, IL-4, IL-6 and IFN-gamma'; 5-LO inhibition 'leading to a decreased production of leukotrienes'; complement suppression 'due to inhibition of the conversion of C3 into C3a and C3b'; prostaglandin inhibition 'appears to play only a minor role'. The bidirectional effects it describes are dose-dependent immune-function endpoints (antibody titres, lymphocyte proliferation), not inflammation. Author's own caveat: which actions 'contribute to the therapeutic effects and which is finally the best dosage of a standardized extract needs further examination.' |
| Zhao W et al 2025 · J Oral Pathol Med | animal | supports | moderate | Rat oral-ulcer model plus in-vitro macrophages (abstract-only; no n, dose or effect sizes reported). In vivo: 'Boswellia extract significantly accelerated oral ulcer healing'; 'Histological examinations revealed reduced tissue swelling and inflammatory cell infiltration in treated groups', with decreased M1 and increased M2 macrophage infiltration. The cytokine and mechanism results are IN VITRO, not in the rat: 'In vitro studies showed that Boswellia extract enhanced M2-related gene expression and decreased pro-inflammatory cytokines, which is PPARγ dependent.' Direction supports the claim; local wound model, not systemic supplement dosing. |
| Kirste 2011 · Cancer | RCT | supports | low | Double-blind placebo-controlled pilot RCT, n=44 irradiated brain-tumour patients, boswellia 4200 mg/day: 'a reduction of cerebral edema of >75% was found in 60% of patients receiving BS and in 26% of patients receiving placebo (P = .023)'. Indirect support only - no inflammatory marker was measured (the endpoint was T2-MRI edema volume, reported as a >75% responder proportion), the authors themselves note 'These findings may be based on an additional antitumor effect', and the clinical secondaries were null: 'The dexamethasone dose during radiotherapy in both groups was not statistically different' and 'BS did not have a significant impact on quality of life or cognitive function'. |
| Mahto K et al 2025 · Inflammopharmacology | observational | supports | low # adjudicated calibration #27 (2026-08-19): no search strategy or risk-of-bias, hedged throughout, own close calls for RCTs to validate — standard narrative-review floor | Narrative review (no new data; no search strategy or risk-of-bias assessment reported) of Boswellia serrata in arthritis management. Asserts boswellic acids 'inhibit primary inflammatory pathways including NFĸB, COX-2, and 5-LOX, thereby reducing inflammation and cartilage degradation,' and that 'Preclinical and clinical studies suggest that Boswellia serrata alleviates pain, improves joint mobility, and slows disease progression.' The review itself concedes the evidence is not settled: 'challenges such as limited large-scale clinical trials, variability in extraction methods, and low bioavailability hinder its clinical application,' and it calls for 'high-quality randomized controlled trials to validate its efficacy and safety.' Zero-weighted (evidence_role: review) — it restates the same Boswellia arthritis trials the hub already counts directly. |
| Mukadam S et al 2026 · Inflammopharmacology | animal | supports | low | Diabetic rat excision-wound model, topical hydrogel, abstract only (PubMed efetch; no open-access full text). The inflammatory finding runs in the claim's direction: 'Topical gel application on the wound significantly reduced the levels of interleukin (IL)-1β, IL-6, tumor necrosis factor alpha (TNF-α), and matrix metalloproteinase-9 (MMP-9) in scraped wound skin on day 7 compared to diabetic controls', and 'These pro-inflammatory cytokines were reduced as the wound healed by day 21, and levels were lower in the treatment group than in the diabetic group.' OFF-SCOPE for this claim on two counts, the same two that already retired the sibling topical burn-wound sources s40430888 and its preprints.org twin: (1) route — the exposure is a gel applied to a wound, not the standardized oral extract at supplement doses this claim scopes itself to; (2) attribution — the cytokine sentences name only 'topical gel application' and 'the treatment group', never the boswellia-alone arm, while the headline formulation is boswellia PLUS zinc, and the paper's own reference list cites zinc as an independent anti-inflammatory and antioxidant (Jarosz 2017; Prasad 2014; Lin 2017). Wound closure IS arm-separated ('Healing effects of BSP + Zn were slightly better than BSP alone (15 mm2, 89.6%), standard gel (13.5 mm2, 91.8%)') but closure is not this claim's outcome, and the boswellia arm did not beat the standard gel. Quality lowered to low: the abstract reports no group sizes, no randomization, allocation or blinding, and no effect size or p-value for any cytokine. Correction to the prior extract — it stated the day-7 cytokine drop came 'alongside faster wound closure', but every closure figure in the paper is day 21; no day-7 closure comparison is reported. Direction of the finding is not disputed; the paper simply does not test the claim as scoped. Mukadam S, Shinde VM, Mahadik KR, Inflammopharmacology 2026;34(5):3655-3669; 'There is no conflict of interest to be declared.' |
| Nischang V et al 2025 · Phytomedicine | in-vitro | supports | low | Human M1/M2 macrophage cultures; two standardized B. serrata extracts differing 10-fold in AKBA (Boswellin Super 30 %, BoswelliaSan 3 %) both evoked 15-LOX product/SPM formation: 'Concentration-response studies revealed only 2-fold increased effectiveness of BSR (30 % AKBA content) versus BOS (3 % AKBA content) for evoking 15-LOX product/SPM formation, and at doses corresponding to 10 uM AKBA each, BOS was much more efficient than BSR.' Conclusion: 'B. serrata extracts promote a beneficial LM class switch in human macrophages, which might not be primarily attributed to AKBA, but to additional compounds.' Mechanism only - no leukotriene, tissue or clinical inflammatory endpoint is reported in the abstract; the leukotriene-suppression wording in the previous extract belonged to the paper's background, not its results. Co-author affiliated with Evonik Operations GmbH despite a 'no conflicts of interest' declaration. |
| Nguyen S, Kaufman M, Shetty M, Rovzar C, Friedlander A, Fredericson M 2025 · Am J Lifestyle Med | observational | supports | low | Narrative (non-systematic) review, PubMed/EMBASE/Cochrane/Google Scholar 2000-2025, no pooling and no new data: 'This narrative review examined the impact of lifestyle interventions on arthritis-related pain and joint health, focusing on osteoarthritis and rheumatoid arthritis.' Its single Boswellia sentence carries no inflammatory-marker result: 'Supplements like curcumin, glucosamine, and Boswellia showed modest benefits with favorable safety' - no effect size, no biomarker, no CI, and Boswellia is bundled with curcumin and glucosamine so its independent contribution is not separable. The abstract's only inflammation finding is assigned to diet, not to Boswellia: 'Anti-inflammatory diets, such as the Mediterranean diet, were linked to lower inflammation and symptom severity.' OFF-SCOPE for this claim on outcome: the outcome reported for the subject is arthritis pain/function benefit, not inflammation. Direction of the reviewed literature is not disputed; the review simply does not test it. Read from the PubMed abstract only - full text not retrievable (fulltext-cache holds a 404 stub; Europe PMC fullTextXML for PMC12618219 returned 0 bytes). |
| Sengupta K, Kolla JN, Krishnaraju AV, Yalamanchili N, Rao CV, Golakoti T, Raychaudhuri S, Raychaudhuri SP 2011 · Mol Cell Biochem | mechanism | supports | low | Manufacturer-authored preclinical package (Laila Impex R&D Center) comparing two of its own standardized B. serrata extracts, Aflapin vs BE-30/5-Loxin. Measured endpoints are anti-inflammatory and all run in the claim's direction: in TNFalpha-stimulated human synovial cells 'the inhibitory potential of Aflapin (IC(50) 44.736 ng/ml) on matrix metalloproteinase-3 (MMP-3) production is 14.83% better than that of BE-30 (IC(50) 52.528 ng/ml)'; 'Aflapin confers better anti-inflammatory efficacy in Freund's Complete Adjuvant (FCA)-induced inflammation model of Sprague-Dawley rats'; and Aflapin 'provides significantly better protection from IL-1beta-induced death of human primary chondrocytes'. Concludes 'both the Boswellia products, BE-30 (5-Loxin) and Aflapin, exhibit powerful anti-inflammatory efficacy'. Supports at preclinical grade only, and weakly: no group sizes, no vehicle-control values and no blinding/randomisation are reported anywhere; the newer in-house product wins on every endpoint with no null; only the abstract is readable (not open access). NOTE the prior extract's 'inhibits 5-lipoxygenase' is not in this paper — the sole '5-Lox' string in the record is the trade name 5-Loxin, and Aflapin is a composition (AKBA-enriched extract plus non-volatile oil fraction), not the molecule AKBA. Mol Cell Biochem 2011;354(1-2):189-197; funding/COI not seen, developer authorship evident from the text. |
| Ammon 2006 · Planta Med | observational | supports | low | Narrative review, one direction only, no counter-evidence reported: 'Animal experiments showed anti-inflammatory activity of the extract... The most evident action is the inhibition of 5-lipoxygenase'; 'Clinical studies, so far with pilot character, suggest efficacy in some autoimmune diseases including rheumatoid arthritis, Crohn's disease, ulcerative colitis and bronchial asthma.' No primary data, no systematic search, single author (group ammon) who is the mechanism's principal proponent - hence review role and low quality. |
| Abdel-Tawab 2011 · Clin Pharmacokinet | observational | supports | low | Narrative review, no new data: 'It underlines BSE as a promising alternative to NSAIDs, which warrants investigation in further pharmacological studies and clinical trials.' Supportive in direction, but the same review undercuts the assumed mechanism — 'boswellic acids failed to inhibit leukotriene formation in human whole blood' and 'the previously assumed mode of action - that is, 5-LO inhibition - is questionable'. Zero-weighted per CONVENTIONS §5. |
| Stengler E et al 2026 · J Ethnopharmacol | in-vitro | supports | moderate | In-vitro only: human airway cells at Air-Liquid-Interface, viral-mimicry and LPS-stimulated models. 'Boswellia serrata exhibited the strongest inhibition of pro-inflammatory cytokines in the viral-mimicry model' - a comparative ranking across eight Burseraceae extracts, with no concentration, effect size, control comparison or p-value given in the available text; the paper tested species oleoresin extracts AND the commercial B. serrata supplement H15 as separate articles and does not say which of the two produced the result. Dosing was bounded only by cytotoxicity ('non-toxic concentrations'), so this speaks to nothing about boswellic-acid levels achievable in human plasma. |
| Anis M et al 2026 · Pak J Pharm Sci | animal | supports | low | Carrageenan-induced paw edema in Wistar albino rats; oral extracts at 400 and 800 mg/kg. Boswellia serrata was one of four single-herb arms and is not reported separately: 'All individual extracts significantly increased pain thresholds and reduced inflammation in carrageenan-induced paw edema assay in dose depended manner as compared to vehicle controls (p < 0.05), with the polyherbal combination producing the highly significant effects (p < 0.001).' No group sizes, randomization, blinding or effect magnitudes are reported, and the extract is unstandardized (no boswellic acid / AKBA content stated). |
| Eid AM et al 2026 · BioMed Research International | in-vitro | supports | low | Cell-free enzyme assay (Cayman COX inhibitor screening kit, Item No. 560131) against bovine COX-1 and human recombinant COX-2 - no cells, no animals, no humans; the anti-inflammatory readout is direct enzyme inhibition. Direction runs with the claim: 'B. sacra oil has IC 50 values of 19.7935 ug/mL on COX-1 and 27.6308 ug/mL on COX-2. As a result, B. sacra oil has anti-inflammatory activity against both COX-1 and COX-2 but is more selective against COX-1. The IC 50 was improved when it was converted to nanoemulgel and became more selective for COX-2, with an IC 50 of 4.3558 ug/mL on COX-1 and 1.3955 ug/mL on COX-2.' Three corrections to the prior extract, which quoted only the 4.36/1.40 nanoemulgel figures as if they were boswellia's. (1) The claim-relevant test article is the NEAT oil, 19.79 / 27.63 ug/mL; the 4.36 / 1.40 numbers belong to a Carbopol gel of a nanoemulsion that is 50% Tween 80 + Span 80 by the paper's own Formulation 1, and no drug-free vehicle was run in this assay (the blank emulgel well existed only in the agar antimicrobial test), so a surfactant contribution is not excluded. (2) Those nanoemulgel values are internally implausible: a diluted formulation cannot be 4.5x (COX-1) and 20x (COX-2) more potent per ug than the neat oil it is made from. (3) Both sets of IC50s are extrapolations outside the measured range - 'The experiment was performed on two occasions using two different concentrations (50 and 300 ug/mL) to ascertain the drugs 50% inhibitory concentration (IC 50) for COX-1/COX-2 activity' - so every reported IC50 (1.3955-27.6308 ug/mL) lies below the lowest concentration actually tested, with no reference inhibitor run alongside. Kept ON-SCOPE, unlike the topical siblings s41910700 / s40430888 / its preprints.org twin: there is no route at all in a cell-free enzyme prep, the neat-oil arm is separable from the topical formulation, and this claim already holds cell-based and nano-formulated in-vitro work on-scope (s41075523, s40876795, d10-21203-rs-3-rs-7302080). Two scope notes for the claim rather than this appraisal: the test article is Boswellia SACRA steam-distilled essential oil (volatile terpene fraction, no AKBA or boswellic-acid content measured in this study), and the enzyme is COX, not the 5-LOX pathway that this claim's note flags as the contested canonical mechanism. Stance, grade and quality unchanged (supports / in-vitro / low). Eid AM et al, BioMed Research International 2026;7034598, doi 10.1155/bmri/7034598. 'Funding No funding was received for this manuscript. Conflicts of Interest The authors declare no conflicts of interest.' |
| Peng C et al 2025 · Front Pharmacol | observational | supports | moderate | Narrative review with a documented multi-database search (PubMed, Web of Science, Scopus, Embase, Google Scholar; stated inclusion/exclusion criteria; no PRISMA counts, no risk-of-bias appraisal, no pooling) and NO data of its own. Supportive bottom line: 'Collectively, preclinical and early-phase clinical trials support the therapeutic potential and favorable safety profile of boswellic acids in various inflammatory disorders' - OA RCTs of Boswel, Aflapin and 5-Loxin 'significantly improve pain scores, joint stiffness, and functional impairment' and 'Some formulations also demonstrated the ability to reduce inflammatory markers (e.g., TNF-alpha, IL-6, CRP) and matrix metalloproteinases (MMPs)'. The review is candid about the other side: 'some RCTs have shown no statistically significant differences between treatment and placebo groups in patients with Crohn's disease or collagenous colitis'; the trials are 'limited by small sample sizes, variable control settings, short follow-up periods, and a lack of biomarker validation'; and 'existing clinical evidence is insufficient to directly validate the standalone clinical efficacy of boswellic acids'. It also concedes the textbook 5-LOX mechanism fails at human exposures: with albumin (10 mg/mL) the inhibition of neutrophil 5-LOX by 11-keto boswellic acid up to 30 uM 'was abolished', which 'explains why' KBA failed to inhibit 5-LOX product formation in human whole blood 'and why single oral administration of Boswellia serrata ... resin extract in a Phase I clinical trial did not suppress plasma leukotriene B4 levels in subjects'. Zero-weighted by evidence_role: review (no new data; the primaries it narrates are already the hub's own sources). |
| Roșca OJ et al 2025 · Preprints.org | animal | supports | low | Preprints.org version of PMID 40430888 (identical 24-rat dataset). Standardized rat second-degree scald burn model: 'Second-degree burns were uniformly induced in 24 Wistar rats using boiling water (100 °C for 8 s) using the RAPID-3D device.' Twenty-one TOPICAL formulations were compared — 'oleogel and hydrogel bases enriched with extracts from Boswellia serrata (frankincense), Ocimum basilicum (basil), Sambucus nigra flower (elderflower), and Galium verum (lady's bedstraw)' — and the reported benefit belongs to a combination arm: 'formulations containing Boswellia serrata and Ocimum basilicum extracts significantly reduced wound size and inflammation, improved skin hydration, and decreased melanin production by days 9 and 21 (p < 0.05).' OFF-SCOPE for this claim on two counts: the route is topical skin application to a burn wound, not the standardized oral extract at supplement doses the claim scopes itself to, and boswellia's independent contribution cannot be separated from co-formulated Ocimum basilicum in the reported arm. Direction of the finding is not disputed. Read from the published version's abstract only (preprints.org returned HTTP 403); full text PMC12114647 not opened. |
| Roșca OJ et al 2025 · Pharmaceutics 2025;17(5):597 | animal | tested-null | low | Rat second-degree scald burn model (24 rats per the abstract, 'Thirty female Wistar rats' per Methods), 21 topical formulations including a Boswellia-alone arm (OG_BS_ETOH70, 5% w/w extract of commercial incense granules; 'Each burn site was treated with 0.2 g of the assigned hydrogel or oleogel formulation'). The Boswellia arm's inflammation outcome was NULL, not positive: at day 21 the significantly lower erythema belonged to the basil arms — 'both melanin (F = 5.11, p < 0.001) and erythema (F = 5.92, p < 0.001) showed significant differences, with oleogel formulations OG_OB_ETOH70 and OG_OB_ETOH99.5 consistently displaying significantly lower values compared to hydrogel formulations (p < 0.05)' — blinded histology 'showed no statistically significant differences between formulations across the evaluated parameters', 'No inflammatory biomarker profiling was performed', and the Boswellia oleogel itself 'exhibited a discrete erythematous halo surrounding the scald lesions'. Extract composition uncharacterized ('characterization ... using HPLC is currently in press'). |
| Mohamed ZA et al 2025 · Research Square | in-vitro | supports | low | Human chondrocyte cell line (C20A4); regular and nano-formulated Boswellia serrata methanol extract significantly suppressed IL1β, IL6, PGE2, and NF-kB inflammation biomarker gene expression. |
| Inferrera F et al 2025 · Neurochem Int | animal | supports | moderate | Reserpine-induced fibromyalgia mouse model; oral Boswellia extract (100 mg/kg) reduced neuroinflammation markers (GFAP, Iba-1 glial activation) and oxidative stress, improving pain/behavioral outcomes. |
| Serafim FGS et al 2026 · Inflammopharmacology. 2026 Jul;34(7):4929-4950 | animal | supports | low | Rat acetic-acid colitis model, BSDE 100 or 300 mg/kg orally for 3 days before and 3 days after induction vs vehicle and dexamethasone 2 mg/kg: 'In the colitis model, BSDE reduced macroscopic damage, preserved crypt architecture, and decreased inflammatory infiltration.' Quality low: abstract-only, no group sizes, no randomization or blinding, and no effect magnitude or p-value is reported for any inflammatory endpoint. The prior extract also placed the oxidative-stress result in the colitis arm; it belongs to the gastric-ulcer arm ('BSDE enhanced gastric mucin content and reduced oxidative stress'). Extract characterized by ESI-MS (six compounds incl. alpha-boswellic and keto-boswellic acid); COX-1/COX-2/iNOS targets are in-silico predictions, not assays. Erratum published (Inflammopharmacology 2026;34(7):4967-4968), content not inspected. |
Disagree, or know a study we missed?
We grade by evidence, not opinions. The way to weigh in is to point us to a study we haven't cited (check the evidence table above first), or to flag a problem with one we have. Every submission is reviewed; if it holds up, the grade updates and shows in Science Changes Its Mind.
Opens a short form. You'll sign in with Google so submissions are tied to a real account — we don't display your identity, and we only accept a link we can verify (PubMed, DOI, ClinicalTrials.gov).
Educational only, not medical advice. Grades and scores reflect published evidence weighted by study design and quality; see the methodology.