The following electronic databases were searched: National Guideline Clearinghouse, Canadian Medical Association Practice Guidelines Infobase, Evidence-based Medicine Guidelines, National Institute for Health and Care Excellence, Royal College of Physicians, The Royal Australian College of General Practitioners, The Cochrane Library, DARE, Bandolier, Medline, TRIP database and Index of Portuguese Medical Journals. The following keywords were applied as search terms: “capsaicin” and “osteoarthritis”.

Guidelines, meta-analysis, systematic reviews and randomized controlled trials (RCTs), published between January 2004 and January 2016, in Portuguese, English or Spanish, were searched. We included articles whose population have been diagnosed with OA, not undergoing arthroplasty. Any form of topical capsaicin (gel, cream, ointment, solution), used alone and compared with a control (placebo or no treatment at all) was included. The measured outcome was the reduction of pain. Secondary outcome included any adverse reaction.

Exclusion criteria were: disagreement with the goal of review; duplicate publication; opinion manuscripts and consensus guidelines; classical review papers and summaries of websites; clinical trials included in recent systematic reviews; and systematic reviews with the same total RCTs as the latest review articles.

Two investigators independently assessed the titles of the articles found and excluded duplicates and those clearly irrelevant. Abstracts of the selected articles were examined independently by the two reviewers who applied the selection criteria. If the information in the abstracts was not enough, full papers were analyzed to make a decision. Where disagreements of selections arose, these were discussed until consensus was reached.

Data were extracted from each eligible study by a single reviewer and checked by the second reviewer. The data items extracted were: number of trials included; number of persons recruited to the trials; type of intervention and control; length of follow-up; evaluated endpoint; results achieved in efficacy and safety; final recommendations.

To evaluate the level of evidence and the strength of recommendation, the Strength of Recommendation Taxonomy (SORT) of the American Academy of Family Physicians was used.17 When the classification was based on other grading scales, we described its meaning.

The systematic review of Cameron et al.18 (Table 1) included thirty-five RCTs that examined the effects of herbal medicinal products in patients diagnosed with OA according to the American College of Rheumatolgy (ACR) criteria. Five RCTs (n=456) compared the analgesic effect of topical capsaicin (0.025% or 0.075%, four times daily) in OA of the hand, knee or multiple joints (hip, knee, shoulder and hand) with placebo. All studies assessed pain (four of them by visual analogue scale – VAS), with similar and consistent results among themselves after three to four weeks of treatment. After three weeks the mean difference (MD) was −18.50% in terms of VAS percentage changes, with a confidence interval (CI) reported at 95% of −40.95 to 3.95. After four weeks, the results were similar (MD −13.90%, CI −32.39 to 4.59). Absolute VAS pain scores after three to four weeks were in favour of capsaicin as well. Thus, topical capsaicin, applied four times daily over three to four weeks reduced the OA pain and was superior to placebo.

Although the placebo cream appear indistinguishable from the capsaicin, placebo validity and blinding was compromised by the local burning sensation associated with topical capsaicin. In one study, this side effect occurred in 44% of participants treated with capsaicin and in only one treated with placebo.

Despite adequate sample size and consistent results for all studies, the level of evidence attributed to this review was two (limited-quality patient-oriented evidence),17 since the allocation concealment was unclear in each of the studies and the blinding was problematic.

In the systematic review of De Silva et al.5 fifty six studies were included to evaluate the evidence regarding complementary and alternative medicine in the treatment of OA (Table 1). Five RCTs (n=427) tested the efficacy of capsaicin gel (0.015%, 0.025% or 0.075%, once daily or four times daily) in the treatment of hand or knee OA compared to placebo. A median Jadad score19 of four was attributed by the authors. The sample sizes ranged from fourteen to two hundred patients and the duration of the studies between four and twelve weeks. In all RCTs it was found that topical capsaicin was significantly more effective than placebo in improving OA pain. After four weeks of treatment, there was a 33% reduction in pain intensity compared with 20% on placebo, whereas in the twelve-week study the reductions were 53 and 27%, respectively. Redness and local burning sensation were the most frequently reported side effects in the capsaicin group (44–46% of patients in two studies).

The main limitations of this study were the small number of trials, the small number of patients in some RCTs and the possibility of publication bias (submission of studies with unfavourable results less likely). The double-blinding method was not assessed in this study. Comparing with the systematic review of Cameron et al.,18 this review included only one new RCT.

Although there is consistent evidence that capsaicin gel is effective in the management of OA, the primary data is of low quality. As such, the level of evidence two was attributed to this review.17

The systematic review of Laslett et al.20 included five RCTs, and one case-crossover trial (n=1162) in which the analgesic effect of topical capsaicin (0.025% or 0.075%, four times daily) was compared with placebo, in the OA of the hand (one study), knee (three studies) or several joints (hand, knee, hip or shoulder) (two studies) (Table 1). Selection criteria varied between trials but patients typically have at least moderate pain and radiological and/or clinical evidence of OA. The sample sizes ranged from fourteen to 695 hundred patients. Most studies applied VAS to assess pain after four to twelve weeks of treatment. It was found that topical capsaicin had a moderate efficacy in reducing pain intensity after four to twelve weeks of treatment, compared to placebo. Indeed, over four weeks of treatment the change in VAS pain score was moderate (standardized mean difference 0.44; 95% CI 0.25–0.62). These results were consistent across trials, suggesting no differences between different doses of capsaicin and between different application sites. It was also reported that continuing use of capsaicin beyond four weeks was beneficial. One study, in a long-term open label extension, reported that differences between groups increase over time, even up to 20 weeks. However, four weeks is probably sufficient to support a therapeutic trial.

Mild localized burning was the most frequently reported adverse event (35–100% of patients affected, with a risk ratio of 4.22, 95% CI 3.25–5.48). It was also reported a greater mean application discomfort score at baseline in the capsaicin group (2.41; 95% CI 1.52–3.30) compared to placebo (0.90; 95% CI 0.21–1.59) (p<0.001). However, burning decreased with continued use. As such, topical capsaicin was considered safe, with no reports of systemic toxicity. Besides, the burning sensation was not associated with clinical response.

The main limitations mentioned in this review were the difficulty of treatment blinding due to the burning sensation associated with capsaicin, the short duration of RCTs and the limited data of capsaicin efficacy in patients with OA at other sites than hand or knee (specially non-superficial joints such as the hip).

Overall, this study suggests that capsaicin should be used for superficial joints like the hand or knee over twelve weeks and possibly longer. The level of evidence two was attributed to this review.17

The guideline from the Royal College of Physicians2 (Table 2) developed its recommendations using studies with high or very high levels of evidence. Four RCTs were found on topical capsaicin given four times daily in OA patients, compared with placebo. These trials evaluated the impact of treatment on pain and morning stiffness, functional capacity and quality of life, with samples of variable size (59–200 patients) and a short study duration (four to twelve weeks). They also differed in terms of osteoarthritis site: knee and shoulder osteoarthritis were evaluated in a RCT; knee osteoarthritis in the second RCT; knee, hip, shoulder and hand osteoarthritis in the third RCT; and hand osteoarthritis in the fourth RCT. It was found that topical capsaicin was superior to placebo in reducing pain intensity (assessed by VAS) and articular tenderness, the difference being statistically significant in most of the variables.

One RCT reported more adverse events in the capsaicin group (69% of patients) compared to placebo (30% of patients). However, no serious side effects have been reported with the use of topical capsaicin.

Thus, based on the evidence, this guideline recommended that topical capsaicin should be considered as an adjunct to core treatment (education, exercise and weight loss) for knee and hand OA with a strength of recommendation B (limited-quality patient-oriented evidence).

The guideline from the Royal Australian College of General Practitioners3 (Table 2) developed recommendations on the non-surgical management of hip and knee OA in primary health care. A single RCT of low quality was found, which included two hundred patients with hip (n=33), knee (n=66), shoulder or hand OA. Topical capsaicin (0.025%, four times daily) was more effective than placebo in improving pain (measured by VAS) after six weeks of treatment.

Its use was associated with local adverse events such as transient local burning and erythema, having higher baseline discomfort scores (averaged over the first 5 days) than other groups. However, this diminishes with continued use. No serious side effects have been reported with the use of topical capsaicin. The authors recommend that the application of capsaicin cream should be with a glove to prevent inadvertent spread to eyes and other mucous membranes.

The main limitation mentioned in this guideline was the small number of participants with OA of the hip and knee in each group, hindering analysis of differences between the groups.

Based on this evidence, the use of topical capsaicin was recommended, alone or in combination with other drugs, for the short-term treatment of mild-moderate persistent symptoms of hip and knee OA, and for acute flares of symptoms. The grade of recommendation D (body of evidence is weak and recommendation must be applied with caution) was attributed.

The guideline from the American College of Occupational and Environmental Medicine21 (Table 2) also issued recommendations supported on clinical evidence and oriented for primary health care. The use of topical capsaicin was recommended for chronic hand OA or acute flares of OA, with a strength of recommendation C (the intervention is recommended, although there is limited evidence regarding the benefits of this intervention).

The authors declare that no experiments were performed on humans or animals for this study.

The authors declare that no patient data appear in this article.

The authors declare that no patient data appear in this article.