Boswellia

Boswellia serrata belongs to Burseraceae, the incense-tree family, and NCBI Taxonomy files it as taxon 613112. Siddiqui (Indian Journal of Pharmaceutical Sciences 2011) describes a moderate to large branching tree of dry mountainous regions; NCCIH gives its native range as the mountainous regions of northern Africa, India, and the Middle East. In India, Siddiqui names Andhra Pradesh, Gujarat, Madhya Pradesh, Jharkhand, and Chhattisgarh as the main sources. LiverTox describes it as a deciduous tree whose paper-thin bark peels away to reveal a gummy oleoresin. Resin is collected by cutting into the trunk; Siddiqui reports that the exudate is stored in bamboo baskets so that the oil drains off and the resin solidifies, after which it is graded by flavour, colour, shape, and size.

Frankincense, or olibanum, is a name shared by several Boswellia species, and the difference matters. Siddiqui notes that the genus has about 25 species, most of them in Arabia, the north-eastern coast of Africa, and India, and that three have traditionally been regarded as true frankincense: Boswellia sacra of southern Arabia, and Boswellia carterii and Boswellia frereana of Somalia. NCBI Taxonomy now treats B. carterii (spelled carteri) as a synonym of B. sacra (taxon 173701). These Arabian and African resins are the classic frankincense of incense, and frankincense essential oils may be sold as B. carterii; the allergy case reports cited by MSKCC, for example, involve B. carterii oil. NCCIH notes that frankincense for aromatherapy may come from B. serrata or other Boswellia species. The oral extracts in most clinical trials are from B. serrata, and Börner and colleagues (Pharmaceuticals 2021) found that frankincense supplements differ strikingly in boswellic acid content, partly because they use material from different species. MSKCC also cautions that boswellia should not be confused with guggul or myrrh, other resins with different chemistry.

Siddiqui reports that the oleo-gum-resin contains 30–60% resin, 5–10% essential oil, and polysaccharides. The resin fraction holds the boswellic acids, pentacyclic triterpenic acids of which four are usually emphasised: beta-boswellic acid (PubChem CID 168928), acetyl-beta-boswellic acid, 11-keto-beta-boswellic acid (KBA, CID 9847548), and acetyl-11-keto-beta-boswellic acid (AKBA, CID 11168203, C32H48O5). Siddiqui and Safayhi’s group both identify AKBA as the most potent inhibitor of 5-lipoxygenase. Commercial extracts are often enriched in particular acids: 5-Loxin, for example, is standardised to 30% AKBA (Sengupta, Arthritis Research and Therapy 2008). An essential oil distilled from the resin captures the volatile fraction and is a different product from the boswellic acid extracts used in the arthritis trials.

Frankincense is one of the oldest traded aromatics. Ernst opens his BMJ review with the gifts of the Magi in Matthew’s gospel, gold, frankincense, and myrrh, and notes that the resin of Boswellia, native to Arabia and India, has a long history in religious ceremonies and perfumery, with medicinal uses appreciated for millennia. Siddiqui writes that Boswellia resin has been used as incense in religious and cultural ceremonies since time immemorial and was prized by the ancient Egyptians, Greeks, and Romans as incense, fumigant, and aromatic. These accounts concern frankincense in general; which species supplied any particular ancient incense is often unclear, so they are not a history of B. serrata specifically.

In India, B. serrata resin, salai or salai guggul, became part of Ayurvedic medicine. Siddiqui notes the Sanskrit name gajabhakshya, suggesting that elephants eat the plant, and reports that the classical Ayurvedic texts describe the antirheumatic action of gugguls, the gum resins of trees. Ayurvedic and Unani texts recommended the resin for a long list of complaints, including diarrhoea, dysentery, skin diseases, fevers, asthma, cough, and jaundice (Siddiqui). MSKCC lists traditional Ayurvedic uses for arthritis, ulcerative colitis, coughs, sores, wound healing, and asthma, and LiverTox notes its use for anti-inflammatory, antiseptic, astringent, and stimulant effects.

Modern pharmacology took off in the late twentieth century. MSKCC cites a 1986 study describing salai guggal extract as a new non-steroidal anti-inflammatory agent that, unlike conventional drugs, showed no pain-relieving or fever-reducing effect. In 1992 Safayhi, Ammon, and colleagues in the pharmacology department of the University of Tübingen reported that boswellic acids are specific inhibitors of 5-lipoxygenase, and Indian and German collaborators went on to publish small trials in ulcerative colitis, asthma, and chronic colitis. Siddiqui notes that the branded extract Boswellin was introduced to US and European markets in 1991. From 2008 onwards, researchers at the Laila Impex R&D Center in Vijayawada, India, published company-funded trials of the enriched extracts 5-Loxin and Aflapin, which feature prominently in the osteoarthritis literature. Boswellia is sold as a dietary supplement; Börner and colleagues note that frankincense preparations have not been approved as a drug.

The main proposed mechanism is inhibition of 5-lipoxygenase. Safayhi and colleagues (Journal of Pharmacology and Experimental Therapeutics 1992) isolated boswellic acids from B. serrata gum resin and found that they reduced the formation of leukotriene B4, an inflammatory signal, in rat white blood cells, with AKBA the most potent. At concentrations up to 400 micromolar they did not impair cyclooxygenase, the target of aspirin and ibuprofen, or 12-lipoxygenase in human platelets, and they did not block the non-enzymatic oxidation of arachidonic acid. The authors concluded that boswellic acids are specific, non-redox inhibitors of 5-lipoxygenase. Leukotrienes drive inflammation in asthma and are thought to sustain inflammation in ulcerative colitis, which is why those diseases were tested early. MSKCC adds preclinical evidence for inhibition of cyclooxygenase-1, the inflammatory regulator NF-kappaB, and production of the cytokine TNF-alpha.

Getting boswellic acids into the blood is the weak link. Sterk and colleagues (Planta Medica 2004) gave healthy men a single dose of a B. serrata extract either fasting or with a high-fat meal, and found that the meal increased blood exposure to beta-boswellic acid, KBA, and AKBA several-fold; some boswellic acids were detectable only after the meal. Products also differ: Börner and colleagues analysed three frequently used frankincense preparations and found striking differences in their anti-inflammatory activity in laboratory assays, which tracked how much boswellic acid each contained. Laboratory potency, absorption, and product quality all have to line up before an effect in patients is plausible.

MSKCC notes that boswellia extracts inhibited platelet aggregation in laboratory studies, attributed to effects on the clotting factors Xa and XIa, and that boswellic acids affected the drug transporters P-glycoprotein, OATP1B3, and MRP2 in vitro. None of these effects has been shown to matter clinically, but they are the basis for the interaction cautions below. Ernst noted no evidence of serious drug interactions in the trials he reviewed, though small, short trials are not designed to detect them.

Knee osteoarthritis is the most studied use. Cameron and Chrubasik’s Cochrane review of oral herbal therapies for osteoarthritis (2014) included 49 trials of many herbs but could pool data only for B. serrata products, from five studies of three different extracts. High-quality evidence from two studies with 85 participants showed that 100 mg a day of an enriched B. serrata extract for 90 days reduced pain by an average of 17 points on a 0–100 scale compared with placebo (from 40 points on placebo), with two people needing treatment for one to benefit, and improved physical function by 8 points. One study with 96 participants suggested fewer adverse events with the extract. Evidence for other boswellia doses and extracts was low or very low quality, and no serious adverse events were reported.

Later reviews point the same way, with the same caveats. Yu and colleagues (BMC Complementary Medicine and Therapies 2020) pooled seven trials with 545 patients and found that boswellia and its extracts reduced pain and stiffness and improved joint function, suggesting treatment for at least four weeks. Bannuru and colleagues (Seminars in Arthritis and Rheumatism 2018) analysed 11 trials of curcumin or boswellia with 1,009 participants: both beat placebo for pain and function, but study quality was low overall, most trials had fewer than 100 participants, no trial compared boswellia with an approved anti-inflammatory drug, and the authors judged the evidence inadequate for clinical recommendations. NCCIH, citing Bannuru’s review among others, says oral boswellia may help osteoarthritis pain and inflammation but that larger, higher-quality studies are needed.

Who ran the trials matters. In the 5-Loxin trial, Sengupta and colleagues (Arthritis Research and Therapy 2008) randomised 75 people with knee osteoarthritis to 100 mg or 250 mg a day of the extract, standardised to 30% AKBA, or placebo for 90 days; both doses improved pain and function, with the higher dose showing benefits as early as day 7. In a follow-up, Sengupta and colleagues (International Journal of Medical Sciences 2010) compared 100 mg a day of 5-Loxin, 100 mg of Aflapin, and placebo in 60 people, of whom 57 finished, and reported that both extracts worked and Aflapin worked better. Both trials were company research: the 2008 paper states that it was funded by the Laila Impex R&D Center, whose employees and consultants wrote it, and the 2010 paper, from the same centre, acknowledges support from Laila Nutraceuticals while declaring no conflict of interest. Ernst’s 2008 review found that funding was undisclosed in all but one trial and that positive publication bias could not be excluded. Kimmatkar’s small crossover trial in 30 patients, included in Ernst’s review and summarised by LiverTox, also reported less knee pain and better walking distance with boswellia.

Inflammatory bowel disease: early studies were small. LiverTox summarises an Indian–German trial in which 42 patients with ulcerative colitis received boswellia gum resin or sulfasalazine for six weeks, with improvement in both groups, and a trial in 30 patients with chronic colitis in which 90% improved on boswellia versus 60% on sulfasalazine. In Crohn’s disease, the German extract H15 was not inferior to mesalazine in a trial of 102 patients, and in a trial of 31 patients with collagenous colitis, 64% went into remission on boswellia versus 27% on placebo, or 44% versus 27% by intention-to-treat analysis (Ernst). The largest and longest trial was negative: Holtmeier and colleagues (Inflammatory Bowel Diseases 2011) randomised 82 people with Crohn’s disease in remission to 2,400 mg a day of the extract Boswelan or placebo for 52 weeks. The trial was stopped early because the drug could not be distinguished from placebo; 59.9% and 55.3% stayed in remission. MSKCC describes the evidence for collagenous colitis as unclear or negative and finds no benefit for maintaining Crohn’s remission.

Asthma and other uses: in an Indian–German trial summarised by Ernst and LiverTox, 80 adults with asthma took boswellia gum resin or placebo for six weeks; 70% improved on boswellia versus 27% on placebo, although Ernst notes that the boswellia group had more severe asthma at the start. NCCIH says a few small studies suggest benefit but the evidence is not rigorous enough to judge. NCCIH also notes a small study in malignant glioma, an aggressive brain tumour, in which oral boswellia appeared to reduce swelling of the brain during treatment but not the tumour itself, and finds too little evidence for topical frankincense in arthritis. NCCIH states there is no evidence that boswellia prevents or treats COVID-19, and MSKCC lists many other small studies, from irritable bowel syndrome to low back pain, that need confirmation.

No regulatory monograph sets a dose. The European Medicines Agency has no herbal monograph for boswellia, and LiverTox notes that it has not been approved for medical uses in the United States. LiverTox gives the usual recommended dose of over-the-counter extracts as 250–500 mg two or three times daily. NCCIH notes that B. serrata extract at up to 1,000 mg daily has been used safely in trials lasting up to six months, and at 2,400 mg daily for up to a month. Ernst reports that 600–3,000 mg of gum resin a day, or equivalent, is usually recommended for oral use. Enriched extracts are dosed much lower: the Cochrane review’s best evidence was for 100 mg a day of an enriched extract for 90 days, and the 5-Loxin trial used 100 or 250 mg a day.

Because products differ so much, the dose on the label means little without knowing what extract it refers to. Siddiqui notes that manufacturing varies from one producer to another, making standardisation difficult, and that the trials used different products from different manufacturers, so their effects may not be comparable. Börner and colleagues found that the anti-inflammatory activity of commercial frankincense preparations tracked their boswellic acid content. Look for a product that names Boswellia serrata, states its content of total boswellic acids or AKBA, and ideally matches an extract tested in a trial. Frankincense essential oil, incense, and resin sold for burning are not substitutes for an oral extract.

Sterk and colleagues found that taking boswellia with a high-fat meal raised blood levels of boswellic acids several-fold, so taking capsules with food is sensible and consistency matters. Yu and colleagues suggested treatment for at least four weeks before judging an effect. Boswellia is often combined with curcumin from turmeric in joint supplements, and Bannuru’s review found both better than placebo, but combination products add each ingredient’s side effects and interactions. Topical frankincense creams and oils have little evidence and can cause allergic skin reactions (NCCIH, MSKCC).