Boswellia serrata for Asthma
Boswellia serrata is not an established treatment for asthma. A few small, dated trials hint at possible breathing benefits, but the evidence is too weak and unreplicated to rely on, and it should never replace prescribed inhalers or controller medication.
Why this grade7-layer evidence engine
This earns a Weak (Grade C) rating because the entire human evidence base is just three small, low-quality randomized trials and no meta-analysis exists. The largest, Gupta 1998 (PMID 9810030, n=80), reported that 70% of the Boswellia group improved versus 27% on placebo, but it had a documented baseline imbalance (the treatment arm started with more severe asthma), reported no p-value, scored only 3/5 on the Jadad scale, and has never been independently replicated in nearly three decades.
The two newer trials do not strengthen the case. A 2015 open-label, industry-context study of a Boswellia phytosome (Casperome) in 32 patients (PMID 26502867) suggested fewer rescue inhalations, but its design is highly prone to bias. A 2018 double-blind trial (PMID 29210124, n=36) showed real symptom and cytokine improvements, yet it tested Boswellia combined with Aegle marmelos, so any benefit cannot be attributed to Boswellia alone.
Authorities reinforce the caution. The FDA treats Boswellia only as a flavoring agent and has issued warning letters against therapeutic claims; the EU/EFSA rejected its health claims and authorizes no human use; and major clinics (Mayo, Cleveland, Harvard) and asthma guidelines (GINA, ATS) are silent on it. The WHO monograph lists asthma as a traditional use 'supported by clinical data,' but that is a low, traditional-medicine bar, not modern proof.
Scoring transparency
All scores computed by a 7-layer evidence engine — fully auditable▸View the full decision path (audit trail)
- compute_raw_score — 加權公式: L2×0.30 + L3×0.25 + L5×0.25 + L11×0.10 + L1×0.10 = 0.53
- tier_from_score — 依分數區間映射至 tier letter
- apply_hec_rules — 無高階證據可裁決
- tier_strict_requirement_check — Tier 條件達標，未降階
- detect_disputes — 偵測到 0 個 hard + 0 個 soft dispute
- decide_status — 依 tier + dispute 結果決定 status