Does comfrey ointment relieve sprain, back and knee pain?

Comfrey root contains allantoin, rosmarinic acid, mucilage, and other compounds, and it is the root that is dug, dried, and extracted into ointments and creams. How these compounds act on an injured joint or muscle has not been worked out. The clinical evidence concerns specific ointments rather than individual ingredients⁴.

How to Use It

The clinical evidence is for standardized topical preparations only: Kytta-Salbe f (a comfrey root extract) in the larger German trials^1,2,3, and Traumaplant (a cream made from the juice of comfrey herb rather than root) in a Czech trial⁷. Homemade comfrey poultices from garden plants have not been tested in these trials, and the strength of their active compounds is unknown.

For ankle sprains: in the head-to-head trial against diclofenac gel, outpatients with an acute sprain applied a 6 cm ribbon of ointment four times a day for about a week. Comfrey was shown to be non-inferior to diclofenac, and the authors say the results indicate it may be superior¹. In a trial of a comfrey herb cream, a 10% cream worked significantly better than a 1% reference cream on pain with movement, pain at rest, and functional impairment at days 3–4 and 7⁷.

For acute low back pain: the trial used 4 g of ointment three times a day for about 5 days³.

For knee osteoarthritis: the 220-patient trial used 2 g of ointment three times a day (6 g a day) for 3 weeks². A smaller trial of 10% and 20% comfrey creams blended with tannic acid used them three times a day for 6 weeks⁵.

Do not apply comfrey to broken skin. The European Medicines Agency's herbal committee concluded that a specific comfrey root extract (applied as a cream or ointment) can be used by adults for the relief of minor sprains and bruises for no longer than 10 days, and that patients should not be exposed to more than 0.35 micrograms of pyrrolizidine alkaloids a day⁸.

See our arnica and boswellia pages for other plant-based topical and oral options for musculoskeletal pain.

The Pyrrolizidine Alkaloid Problem

Comfrey contains as many as 14 pyrrolizidine alkaloids (PAs), including symphytine, echimidine, intermedine, and lycopsamine. The available evidence suggests that the active metabolites of these alkaloids interact with DNA in liver cells, causing DNA damage, mutations, and cancer. Comfrey is toxic to the liver in livestock and humans and causes cancer in experimental animals⁶. Because PAs have toxic effects on the liver when swallowed⁸, comfrey should never be taken by mouth. In 2001 the U.S. FDA advised manufacturers to remove oral comfrey products from the market.

The European Medicines Agency's committee considers the risk from comfrey root medicines used on the skin for short periods to be not significant, but it requires that the PA content of each product be specified and that exposure stay below 0.35 micrograms a day⁸.

The bottom line: never take comfrey by mouth, never apply it to broken skin, and keep to short courses of commercial ointments that state their PA content rather than homemade preparations from garden plants.

The clinical case for topical comfrey rests on a small set of randomized trials. The three largest tested the same branded ointment (Kytta-Salbe f), and their authors overlap (Predel, Giannetti, Staiger, Bulitta)^1,2,3.

Predel et al. 2005 (Phytomedicine) is the active-comparator trial (called CODEC_2004): 164 outpatients (mean age 29.0, 47.6% women) with an acute unilateral ankle sprain were randomized to a 6 cm layer of Kytta-Salbe f comfrey extract ointment (n = 82) or to diclofenac gel containing 1.16 g of diclofenac diethylamine salt (n = 82), four times a day for 7 ± 1 days. The primary outcome was the area under the curve (AUC) of the pain reaction to pressure on the injured area, measured with a calibrated caliper (tonometer). It was confirmed that comfrey was non-inferior to diclofenac: the 95% confidence interval for the AUC difference (comfrey minus diclofenac) was 19.01 to 103.09 h·N/cm², completely above the non-inferiority margin. The authors added that the primary and secondary results indicate comfrey may be superior to diclofenac gel. Secondary outcomes were swelling, pain at rest and on movement, joint mobility, use of rescue paracetamol, and global assessments of efficacy and tolerability.¹

Grube et al. 2007 (Phytomedicine) is the placebo-controlled knee osteoarthritis trial: 220 patients (153 women, 67 men, average age 57.9 years; complaints for 6.5 years on average) were randomized to Kytta-Salbe f or placebo, 2 g three times a day for 3 weeks. The visual analog scale (VAS) total score, the primary outcome, fell by 51.6 mm (54.7%) in the comfrey group and by 10.1 mm (10.7%) with placebo, a difference of 41.5 mm (95% CI 34.8 to 48.2 mm, p < 0.001). The WOMAC total score fell by 60.4 mm (58.0%) with comfrey and by 14.7 mm (14.1%) with placebo, a difference of 45.7 mm (95% CI 37.1 to 54.3 mm, p < 0.001). Quality of life (SF-36), knee mobility, and the global assessments by physician and patient also favored comfrey (p < 0.001 each).²

Giannetti et al. 2010 (British Journal of Sports Medicine) tested topical comfrey for acute upper or lower back pain in 120 adults (mean age 36.9) randomized to comfrey root extract ointment or placebo, 4 g three times a day for about 5 days. The primary outcome was the area under the curve of VAS pain on active standardized movement. Pain on active movement fell by a median of about 95.2% with comfrey and 37.8% with placebo, a significant difference. The results were consistent across all primary and secondary outcomes (pain at rest, pressure algometry, global efficacy assessments, analgesic use, and the Oswestry disability index), and for the first time a fast-acting effect of the ointment (within 1 hour) was seen.³

Kucera et al. 2004 (Wiener Medizinische Wochenschrift) tested Traumaplant, a cream containing 10% of a pressed juice of fresh comfrey herb (Symphytum x uplandicum), against a 1% reference cream in 203 patients with acute ankle distortion (104 high-concentration, 99 low-concentration). With the high concentration, scores for pain on active motion, pain at rest, and functional impairment fell significantly on days 3–4 and day 7 (p < 0.001), and swelling improved significantly on days 3–4 (p < 0.01). Efficacy was judged good to excellent in 85.6% of cases with the high concentration and 65.7% with the reference cream on days 3–4, and overall tolerability was excellent. The comparator was a weaker version of the same cream, not a placebo or an NSAID.⁷

Smith and Jacobson 2011 (Journal of Chiropractic Medicine) enrolled 43 people aged 45–83 with primary knee osteoarthritis and randomized them to a 10% comfrey root extract cream, a 20% cream, or a placebo cream; all three also contained eucalyptus, and the comfrey creams contained a blend with tannic acid. They applied the cream three times a day for 6 weeks. Pain, stiffness, and daily function on the WOMAC index were all significantly better with both comfrey creams than with the reference cream (p < 0.01 for each). Two participants in each active group had temporary, minor skin rash or itching. It was a small trial and used a blend of ingredients, so it cannot isolate the effect of comfrey.⁵

Staiger 2012 (Phytotherapy Research) is a clinical overview of the trials and non-interventional studies published on comfrey to that date. It concludes that multiple randomized controlled trials have demonstrated the efficacy and safety of comfrey preparations for the topical treatment of pain, inflammation, and swelling of muscles and joints in degenerative arthritis, acute back myalgia, sprains, contusions, and strains, including in children aged 3 or 4 and over.⁴

Mei et al. 2010 (Journal of Toxicology and Environmental Health, Part B) reviewed comfrey's toxicity. Comfrey has been consumed as a vegetable, tea, and herbal medicine for more than 2000 years, but it produces hepatotoxicity in livestock and humans and carcinogenicity in experimental animals. It contains as many as 14 pyrrolizidine alkaloids. The review discusses evidence that active PA metabolites interact with DNA in liver endothelial cells and hepatocytes, causing DNA damage, mutation induction, and cancer; comfrey and the model PA riddelliine produce similar DHP-derived DNA adducts and similar mutation spectra. It concludes that comfrey is mutagenic in the liver and that its PAs appear responsible for the toxicity and tumors.⁶

The EMA HMPC summary⁸ is the regulatory view. The committee concluded, on the basis of long-standing use, that a specific ethanol-extract comfrey root preparation in a semi-solid form can be used for the relief of symptoms of minor sprains and bruises by adults for no longer than 10 days. This is a "traditional use" conclusion: the committee said there is insufficient evidence from clinical trials to support it. It noted four clinical studies with a different comfrey preparation that suggested reduced swelling and pain but did not take them into account because the exact composition of that preparation is unknown. No side effects had been reported with the covered medicines at the time of the assessment, and no significant risk was expected when used on the skin for short periods. Patients should not be exposed to more than 0.35 micrograms of PAs per day.

Strength of evidence: moderate for topical comfrey extract ointments in acute soft-tissue injury, knee osteoarthritis, and acute back pain, based on several randomized trials with consistent results. The main limitations are that the three largest trials tested one branded ointment and share authors, the follow-up is short (5 days to 6 weeks), the comparator in the ankle trials was either an NSAID gel or a weaker version of the same cream, and there is no long-term safety surveillance for chronic topical use. The European regulator's own conclusion rests on traditional use, with short courses and a cap on PA exposure. Internal use of comfrey is not safe and should be avoided.