Osteoarthritis and Joint Pain

Osteoarthritis is the flagship use for devil's claw, and it is the one where the trials are longest. The centrepiece is a four-month, multicentre, double-blind randomised trial in 122 people with knee or hip osteoarthritis, in which devil's claw was compared not with a sugar pill but with a prescription osteoarthritis drug. That is a serious study design, and it produced a result worth knowing about. It also has a structural limitation that most write-ups skip over, and this article does not.

Table of Contents

  1. What Osteoarthritis Is, and What a Herb Can Do About It
  2. The Four-Month Trial Against Diacerein
  3. The Nonsteroidal Anti-Inflammatory Drug Question — Diclofenac and Rofecoxib, Stated Precisely
  4. The Standardised-Extract Studies — 250 Patients and 75 Patients
  5. The 259-Patient Open Study, and What “Open” Costs You
  6. What the Systematic Reviews Concluded
  7. Why It Might Work — the Laboratory Story
  8. The Doses Actually Used in the Osteoarthritis Trials
  9. How People Track an Eight-to-Twelve Week Trial
  10. Cautions That Matter Specifically in Osteoarthritis
  11. The Bottom Line
  12. Key Research Papers
  13. Connections
  14. Featured Videos

What Osteoarthritis Is, and What a Herb Can Do About It

Osteoarthritis is not simply cartilage wearing out. It is a whole-joint disease: the cartilage thins and fissures, the bone underneath thickens and grows spurs, the joint lining (synovium) becomes intermittently inflamed, ligaments loosen, and the muscles around the joint weaken. The pain comes partly from that low-grade synovial inflammation and partly from bone, capsule and nerve changes — which is why the amount of damage on an X-ray correlates so poorly with how much a person hurts.

That matters for what to expect from devil's claw. A plant extract with anti-inflammatory activity has a plausible target: the inflammatory component of the pain. It has no plausible route to regrowing cartilage, and nobody has shown it does. Every honest claim about devil's claw in osteoarthritis is a claim about symptoms — pain, stiffness, how far you can walk, how much paracetamol or ibuprofen you end up taking — over weeks to months.

If you are being told devil's claw “rebuilds joints,” “reverses arthritis” or “regenerates cartilage,” the seller has left the evidence behind. There is no imaging trial, no joint-space-width endpoint, no biomarker study supporting any of that.

The Four-Month Trial Against Diacerein

The single most important osteoarthritis study on devil's claw was run by a French group and published twice in 2000 — once in a phytotherapy journal and once in a rheumatology journal, with slightly different emphasis. The design:

What it found

Both groups improved significantly and progressively over the four months, on pain, on functional disability and on the Lequesne index — and there was no statistically significant difference between them on any efficacy measure. Devil's claw held its own against the drug.

Two secondary findings were arguably more interesting than the primary one. By the end of the study, patients in the devil's claw group were using significantly fewer rescue analgesics and non-steroidal anti-inflammatories than the diacerein group. And adverse events were significantly less frequent on devil's claw: diarrhoea, the commonest complaint in both arms, occurred in 8.1 percent of the devil's claw group versus 26.7 percent of the diacerein group. Patients' own global tolerance rating favoured devil's claw.

The limitation nobody mentions

Here is the structural problem: there was no placebo arm. Osteoarthritis pain is famously responsive to placebo — meta-analyses of osteoarthritis trials routinely find large placebo responses, and a four-month study with regular clinic visits and attentive investigators is exactly the setting that maximises them. Both groups improving substantially is therefore consistent with two very different stories:

  1. Devil's claw and diacerein are both genuinely effective and roughly equal, or
  2. Neither did much beyond what an inert capsule plus four months of attention would have done.

This trial cannot distinguish those. What it establishes solidly is non-inferiority to diacerein with better tolerability — which is a real and useful finding, especially because diacerein itself has been repeatedly questioned and restricted in Europe over efficacy and diarrhoea. Being as good as a modest drug is not the same as being proven effective. Say that out loud and the literature becomes much easier to read.

The Nonsteroidal Anti-Inflammatory Drug Question — Diclofenac and Rofecoxib, Stated Precisely

Devil's claw is often described as “tested against diclofenac and rofecoxib.” Two different things are being blurred together there, and the distinction is worth getting right.

ComparisonWhat actually happenedWhat it supports
Diacerein (osteoarthritis, four months, n = 122)A genuine randomised double-blind head-to-head. No difference in efficacy; fewer adverse events on devil's clawNon-inferiority to diacerein in knee and hip osteoarthritis
DiclofenacDiclofenac was the permitted rescue nonsteroidal anti-inflammatory drug (NSAID) in the diacerein trial, not a randomised comparator. Devil's claw patients used significantly less of itAn analgesic-sparing effect — suggestive, and a real patient-relevant outcome, but not a head-to-head against diclofenac
Rofecoxib (low back pain, six weeks, n = 88)A genuine randomised double-dummy head-to-head — but explicitly a pilot, in back pain rather than osteoarthritis, and underpoweredNothing conclusive. The authors' own conclusion was that “large numbers will be needed to show equivalence”

So: there is no adequately powered randomised trial of devil's claw against diclofenac, or against ibuprofen, or against naproxen, in osteoarthritis. Anyone who tells you devil's claw “matched diclofenac” is over-reading a rescue-medication count. The rofecoxib pilot is covered properly in the low back pain article, because that is the condition it studied.

One small randomised trial has since compared devil's claw head-to-head with a different nonsteroidal anti-inflammatory drug, meloxicam. In that trial, 60 people aged 40–60 with painful knee osteoarthritis of grades 1–2 on the Kellgren-Lawrence scale (the milder end) took either two devil's claw tablets a day (2 × 480 mg) for one month or meloxicam 15 mg a day for ten days. Pain on a visual analogue scale (VAS), WOMAC and Oxford Knee Scale scores improved in both groups over time, but neither treatment was significantly better: at 8 weeks, VAS pain was 4.80 with devil's claw versus 5.06 with meloxicam, and WOMAC 25.73 versus 26.20. The authors called devil's claw effective and safe in mild knee osteoarthritis in the short term, and noted that the two treatment periods differed.

The analgesic-sparing signal is still worth something. If you take a nonsteroidal anti-inflammatory drug daily for knee osteoarthritis, needing less of it is a genuinely valuable outcome — nonsteroidal anti-inflammatory drugs (NSAIDs) carry gastrointestinal bleeding, kidney and cardiovascular risks that accumulate with dose and duration, and those risks fall as the dose falls.

The Standardised-Extract Studies — 250 Patients and 75 Patients

A German group ran a series of post-marketing surveillance studies on a standardised proprietary aqueous devil's claw extract. These are not randomised trials, and the authors did not pretend otherwise, but they are large and they report their dose properly.

The 250-patient study

250 patients with either non-specific low back pain (n = 104), knee osteoarthritis (n = 85) or hip osteoarthritis (n = 61) took an eight-week course of the extract providing 60 mg of harpagoside per day. 227 completed. Every outcome measure — validated ones like the WOMAC index and the Arhus index, plus generic quality-of-life measures — improved by week four and improved further by week eight.

Two details from this study are genuinely informative. The hip group improved more convincingly than the back group; the knee group sat in between. And among the back-pain patients who needed nonsteroidal anti-inflammatory drugs during the eight weeks, nonsteroidal anti-inflammatory drug use declined with time. Roughly 10 percent of patients had minor adverse events possibly attributable to the extract. The authors' summary was that between 50 and 70 percent of patients benefited.

That last figure needs a translation. “50 to 70 percent benefited” in an open study is not the same as “50 to 70 percent benefited because of the herb.” With no control group, the natural history of an osteoarthritis flare, regression to the mean, and expectation are all folded into that number.

The 75-patient study

A separate 12-week multicentre surveillance study enrolled 75 patients with hip or knee arthrosis and gave 2,400 mg of aqueous extract daily, corresponding to 50 mg of harpagoside. The WOMAC subscales fell by 23.8 percent for pain, 22.2 percent for stiffness and 23.1 percent for physical function; the total WOMAC index fell 22.9 percent. Visual-analogue pain scores dropped by 22–26 percent depending on which pain question was asked. Physicians recorded improvement in pain on palpation (45.5 percent), limitation of mobility (35 percent) and joint crepitus (25.4 percent). Only two possible adverse drug reactions were reported, both digestive: dyspepsia and a sensation of fullness.

A 23 percent fall in WOMAC is a clinically meaningful size — it is roughly what you would hope for from a nonsteroidal anti-inflammatory drug. But again: uncontrolled, open, and industry-adjacent. The authors said so themselves in their conclusion, writing that the results “suggest” a beneficial effect.

The 259-Patient Open Study, and What “Open” Costs You

The largest devil's claw study in a British population enrolled 259 patients with arthritis and other rheumatic conditions in a single-group open study lasting eight weeks. It measured global pain, stiffness and function on numeric rating scales, plus the WOMAC index and a hand-osteoarthritis index, with safety assessed by self-report, blood analysis and liver function tests.

Results were strongly positive: statistically significant improvements (p < 0.0001) in global pain, stiffness and function; significant reductions in pain scores for hand, wrist, elbow, shoulder, hip, knee and back; significant improvement in SF-12 quality of life; and — the headline that gets quoted everywhere — 60 percent of patients either reduced or stopped their concomitant pain medication.

Now the honest reading. A single-group open study has no comparison. Every patient knew what they were taking, every patient had signed up hoping it would help, and every patient was being seen and assessed regularly for eight weeks. The p-value of 0.0001 tells you the improvement was very unlikely to be chance variation; it tells you nothing about whether the herb caused it. A placebo arm in the same study would almost certainly also have improved significantly.

What the study does contribute solidly is safety data in a decent-sized real-world population, including bloods and liver function — and there the news was reassuring. Read it as a tolerability study with encouraging efficacy signals, not as proof of effect.

What the Systematic Reviews Concluded

Reviewers who pooled this literature converged on a similar verdict, and it has barely moved in twenty years.

The most telling verdict is the regulatory one. As set out on the Benefits hub, the European Medicines Agency's herbal committee reviewed this entire body of work and kept devil's claw in the traditional use category rather than promoting it to well-established use — declining an explicit industry request to reclassify it during the 2016 revision. If the osteoarthritis evidence had been strong enough, that is where it would have shown.

Why It Might Work — the Laboratory Story

The mechanistic work is genuinely interesting, and it is genuinely preclinical.

The gap between all that and a human knee is the concentration problem. Cell studies use concentrations achieved by pipetting extract into a well. Whether swallowing a capsule produces anything like those concentrations in synovial fluid has not been demonstrated, and harpagoside's behaviour in the acidic stomach complicates the question further — see the dosing article. “Devil's claw is anti-inflammatory” is a well-supported hypothesis, not an established fact about people.

The Doses Actually Used in the Osteoarthritis Trials

StudynPreparationDaily doseDuration
Diacerein comparison122Cryoground root powder capsules2,610 mg powder (6 × 435 mg)4 months
Standardised-extract surveillance250 (146 with knee or hip osteoarthritis)Standardised aqueous extract60 mg harpagoside8 weeks
Aqueous-extract surveillance75Aqueous extract2,400 mg extract = 50 mg harpagoside12 weeks
UK open study259Proprietary devil's claw tabletsper manufacturer's label8 weeks

Notice the convergence: where harpagoside was reported, it was 50 to 60 mg per day. That band is the dose the trials converged on. Turning any given label into that number is arithmetic, and the dosing article works it through step by step.

How People Track an Eight-to-Twelve Week Trial

Devil's claw is cheap, generally well tolerated and slow-acting, and the trials above give a template for judging it. People who try it in a structured way, with the cautions below in view, commonly track it like this:

  1. Choosing by harpagoside, not by price or milligrams of root. The trials that reported harpagoside used 50–60 mg a day. A product that states its harpagoside content lets the number of units per day be worked out; a label that does not state it leaves that sum impossible.
  2. A baseline written down before starting. Three numbers, each 0–10: worst pain in the last week, average pain in the last week, and how much the joint limits daily activity — plus one function marker that can actually be measured: how many stairs, how far a walk goes before stopping, whether a jar can be opened.
  3. A record of other painkillers. A week's count of paracetamol and nonsteroidal anti-inflammatory drug tablets before starting. Reduced painkiller use was one of the most consistent findings in the trials, and it is easier to notice than a two-point pain change.
  4. Eight weeks at least, twelve where possible. The surveillance studies showed improvement at four weeks and more at eight. A judgement at day five measures the wrong thing.
  5. One change at a time. Starting devil's claw, a new exercise programme and turmeric in the same week makes it impossible to tell which did what.
  6. Food and the stomach. Digestive upset is the main reason people quit, and people commonly take it with food for that reason.
  7. A second score at week eight, against the same numbers. When nothing has moved — pain, function, or tablet count — that is the answer for that person. Devil's claw does not work for everybody, and paying for a supplement that is doing nothing is a cost of its own.
  8. The treatments with the strongest evidence. For knee and hip osteoarthritis, the interventions with the strongest evidence are exercise therapy, strength training around the joint and weight management. No herb replaces those. Devil's claw is at best a useful addition at the margin.

Cautions That Matter Specifically in Osteoarthritis

The full safety discussion is in the safety, interactions and sustainability article. Three of those cautions are disproportionately relevant to people with osteoarthritis, because of who tends to have it.

The standard cautions given for devil's claw also apply: active peptic ulcer disease, gastritis and gallstones, and pregnancy, along with stopping it a week or two before planned surgery and telling the surgical team.

The Bottom Line

For knee and hip osteoarthritis, devil's claw has the best evidence base of almost any single herb sold for the condition — a four-month randomised double-blind trial showing it matched a prescription drug with fewer side effects, plus several hundred patients' worth of open surveillance data showing WOMAC improvements around 20 percent and substantial reductions in painkiller use.

And: the pivotal trial had no placebo arm, the supporting studies were open and uncontrolled, no adequately powered head-to-head against a common nonsteroidal anti-inflammatory drug exists, the European regulator declined to promote it out of the traditional-use category, and the effect — where real — is modest and symptomatic.

Where harpagoside was reported, the trials used 50–60 mg a day, and reduced use of anti-inflammatory tablets was one of their most consistent findings. Devil's claw is not a substitute for exercise therapy, weight management or a proper assessment of the joint. And if it has not moved any of a person's numbers by week twelve, it is not going to.

Key Research Papers

Every identifier below was verified live against NCBI E-utilities — first author, title, journal and year all had to match before the number was written.

The randomised comparison trials

  1. Chantre P, Cappelaere A, Leblan D, Guedon D, Vandermander J, Fournie B. Efficacy and tolerance of Harpagophytum procumbens versus diacerhein in treatment of osteoarthritis. Phytomedicine. 2000;7(3):177–183. — PubMed PMID: 11185727
  2. Leblan D, Chantre P, Fournié B. Harpagophytum procumbens in the treatment of knee and hip osteoarthritis. Four-month results of a prospective, multicenter, double-blind trial versus diacerhein. Joint Bone Spine. 2000;67(5):462–467. — PubMed PMID: 11143915
  3. Chrubasik S, Model A, Black A, Pollak S. A randomized double-blind pilot study comparing Doloteffin and Vioxx in the treatment of low back pain. Rheumatology (Oxford). 2003;42(1):141–148. — PubMed PMID: 12509627
  4. Farpour HR, Rajabi N, Ebrahimi B. The Efficacy of Harpagophytum procumbens (Teltonal) in Patients with Knee Osteoarthritis: A Randomized Active-Controlled Clinical Trial. Evidence-Based Complementary and Alternative Medicine. 2021;2021:5596892. — PubMed PMID: 34712343

Open and surveillance studies

  1. Chrubasik S, Thanner J, Künzel O, Conradt C, Black A, Pollak S. Comparison of outcome measures during treatment with the proprietary Harpagophytum extract Doloteffin in patients with pain in the lower back, knee or hip. Phytomedicine. 2002;9(3):181–194. — PubMed PMID: 12046857
  2. Wegener T, Lüpke NP. Treatment of patients with arthrosis of hip or knee with an aqueous extract of devil's claw (Harpagophytum procumbens DC.) Phytotherapy Research. 2003;17(10):1165–1172. — PubMed PMID: 14669250
  3. Warnock M, McBean D, Suter A, Tan J, Whittaker P. Effectiveness and safety of devil's claw tablets in patients with general rheumatic disorders. Phytotherapy Research. 2007;21(12):1228–1233. — PubMed PMID: 17886223
  4. Chrubasik S, Chrubasik C, Künzel O, Black A. Patient-perceived benefit during one year of treatment with Doloteffin. Phytomedicine. 2007;14(6):371–376. — PubMed PMID: 17521896

Reviews and critical appraisal

  1. Gagnier JJ, Chrubasik S, Manheimer E. Harpagophytum procumbens for osteoarthritis and low back pain: a systematic review. BMC Complementary and Alternative Medicine. 2004;4:13. — PubMed PMID: 15369596
  2. Brien S, Lewith GT, McGregor G. Devil's claw (Harpagophytum procumbens) as a treatment for osteoarthritis: a review of efficacy and safety. Journal of Alternative and Complementary Medicine. 2006;12(10):981–993. — PubMed PMID: 17212570
  3. Chrubasik S, Conradt C, Black A. The quality of clinical trials with Harpagophytum procumbens. Phytomedicine. 2003;10(6–7):613–623. — PubMed PMID: 13678253
  4. Chrubasik JE, Roufogalis BD, Chrubasik S. Evidence of effectiveness of herbal antiinflammatory drugs in the treatment of painful osteoarthritis and chronic low back pain. Phytotherapy Research. 2007;21(7):675–683. — PubMed PMID: 17444576
  5. Denner SS. A review of the efficacy and safety of devil's claw for pain associated with degenerative musculoskeletal diseases, rheumatoid, and osteoarthritis. Holistic Nursing Practice. 2007;21(4):203–207. — PubMed PMID: 17627199
  6. Menghini L, Recinella L, Leone S, et al. Devil's claw (Harpagophytum procumbens) and chronic inflammatory diseases: a concise overview on preclinical and clinical data. Phytotherapy Research. 2019;33(9):2152–2162. — PubMed PMID: 31273865
  7. Hong JY, Lee J, Kim H, Kim H, Jeon WJ, Yeo C, Lee YJ, Go HY, Ha IH. Harpagophytum procumbens in musculoskeletal disorders: current evidence and comparison with NSAIDs. Frontiers in Pharmacology. 2026;17:1839470. — PubMed PMID: 42394971

Mechanism — laboratory and animal work

  1. Fiebich BL, Muñoz E, Rose T, Weiss G, McGregor GP. Molecular targets of the antiinflammatory Harpagophytum procumbens (devil's claw): inhibition of TNFα and COX-2 gene expression by preventing activation of AP-1. Phytotherapy Research. 2012;26(6):806–811. — PubMed PMID: 22072539
  2. Fiebich BL, Heinrich M, Hiller KO, Kammerer N. Inhibition of TNF-alpha synthesis in LPS-stimulated primary human monocytes by Harpagophytum extract SteiHap 69. Phytomedicine. 2001;8(1):28–30. — PubMed PMID: 11292236
  3. Huang TH, Tran VH, Duke RK, et al. Harpagoside suppresses lipopolysaccharide-induced iNOS and COX-2 expression through inhibition of NF-kappaB activation. Journal of Ethnopharmacology. 2006;104(1–2):149–155. — PubMed PMID: 16203115
  4. Grant L, McBean DE, Fyfe L, Warnock AM. The inhibition of free radical generation by preparations of Harpagophytum procumbens in vitro. Phytotherapy Research. 2009;23(1):104–110. — PubMed PMID: 18803229
  5. Lanhers MC, Fleurentin J, Mortier F, Vinche A, Younos C. Anti-inflammatory and analgesic effects of an aqueous extract of Harpagophytum procumbens. Planta Medica. 1992;58(2):117–123. — PubMed PMID: 1529021
  6. Kaur H, Kumar D, Gauttam VK, et al. Decoding the mechanistic landscape of harpagoside: from molecular targets to translational pharmacology. Fitoterapia. 2026;188:107029. — PubMed PMID: 41391813

PubMed Topic Searches

  1. PubMed: Harpagophytum and osteoarthritis
  2. PubMed: Devil's claw and knee pain
  3. PubMed: Harpagophytum and WOMAC index
  4. PubMed: Harpagoside and COX-2
  5. PubMed: Diacerein in osteoarthritis
  6. PubMed: Herbal medicine for osteoarthritis, randomised trials
  7. PubMed: Placebo response in osteoarthritis trials
  8. PubMed: Exercise therapy for knee osteoarthritis

Connections


Back to Table of Contents