Patient-friendly summary
If you read nothing else
Bottom line
Topical NSAIDs are effective and safe for chronic musculoskeletal pain over two weeks, performing as well as oral NSAIDs, but longer and larger trials are needed to confirm their long-term role.
Moderate evidencePublished
Evidence hierarchy
Study participants
Adults, generally over 40, with moderate-to-severe chronic musculoskeletal pain, predominantly osteoarthritis (mostly of the knee)
Study Summary
This systematic review and meta-analysis pooled 25 randomised, double-blind trials of topical NSAIDs (gels, creams, sprays) for chronic musculoskeletal pain, mostly osteoarthritis. Across 14 placebo-controlled trials (1,502 patients), topical NSAIDs were significantly better than placebo at achieving roughly 50% pain relief at two weeks, with a number-needed-to-treat (NNT) of 4.6 — meaning about one in every four to five treated patients benefited who would not have with placebo. Three trials (764 patients) comparing topical with oral NSAIDs found no difference in efficacy, and topical NSAIDs were as safe as placebo for local and systemic side effects over the two-week period. The authors note the trials were short and often small, so longer, larger studies are needed to confirm the role of topical NSAIDs in practice.
Key Findings
| Finding | Detail | Impact |
|---|---|---|
| Topical NSAIDs beat placebo for chronic musculoskeletal pain at two weeks | Across 14 placebo-controlled trials (1,502 patients), relative benefit was 1.9 (95% CI 1.7 to 2.2) and NNT was 4.6 (95% CI 3.8 to 5.9) for at least ~50% pain relief at two weeks. Mean response was 48% with topical NSAID versus 26% with placebo. | High |
| No efficacy difference between topical and oral NSAIDs | Three trials (764 patients) in knee or finger osteoarthritis found a 37% successful outcome with both topical and oral NSAID, with no statistically significant difference (relative risk 1.1; 95% CI 0.9 to 1.3). | High |
| Topical NSAIDs were as safe as placebo over two weeks | In placebo-controlled trials there was no significant difference between topical NSAID and placebo for local adverse events (6%), systemic adverse events (3%), or withdrawals due to adverse events (1%). Local effects were usually mild rash, itching or stinging. | High |
| Result was robust across sensitivity analyses | Efficacy was not affected by trial quality, validity, size, the outcome reported, or condition treated (knee osteoarthritis versus other musculoskeletal). The 10 highest-quality/validity trials had a similar NNT of 4.4 (95% CI 3.6 to 5.6). | Medium |
| Efficacy estimate worsened compared with the earlier review | After removing lower-quality and non-double-blind trials and reclassifying salicylate and benzydamine as non-NSAIDs, the NNT rose (worsened) from the prior 3.1 (95% CI 2.7 to 3.8) to about 4.6-4.7. | Medium |
| Local side effects slightly higher than oral NSAID, but not systemic harm | In two active-controlled trials, local adverse events were more frequent with topical (8%) than oral NSAID (3%); systemic adverse events and withdrawals did not differ. No upper GI bleeding or symptomatic ulcers were documented in any study. | Medium |
Across 14 placebo-controlled trials (1,502 patients), relative benefit was 1.9 (95% CI 1.7 to 2.2) and NNT was 4.6 (95% CI 3.8 to 5.9) for at least ~50% pain relief at two weeks. Mean response was 48% with topical NSAID versus 26% with placebo.
Three trials (764 patients) in knee or finger osteoarthritis found a 37% successful outcome with both topical and oral NSAID, with no statistically significant difference (relative risk 1.1; 95% CI 0.9 to 1.3).
In placebo-controlled trials there was no significant difference between topical NSAID and placebo for local adverse events (6%), systemic adverse events (3%), or withdrawals due to adverse events (1%). Local effects were usually mild rash, itching or stinging.
Efficacy was not affected by trial quality, validity, size, the outcome reported, or condition treated (knee osteoarthritis versus other musculoskeletal). The 10 highest-quality/validity trials had a similar NNT of 4.4 (95% CI 3.6 to 5.6).
After removing lower-quality and non-double-blind trials and reclassifying salicylate and benzydamine as non-NSAIDs, the NNT rose (worsened) from the prior 3.1 (95% CI 2.7 to 3.8) to about 4.6-4.7.
In two active-controlled trials, local adverse events were more frequent with topical (8%) than oral NSAID (3%); systemic adverse events and withdrawals did not differ. No upper GI bleeding or symptomatic ulcers were documented in any study.
Strengths
- Followed QUOROM guidelines with independent, multi-reviewer quality (Jadad) and validity scoring and dispute resolution by discussion
- Comprehensive search across multiple databases and direct requests to 88 pharmaceutical companies worldwide for unpublished data
- Stricter inclusion criteria than the prior review (double-blind only, minimum trial size) and reclassification of non-NSAID topicals improved rigour
- Robust result confirmed across multiple pre-planned sensitivity analyses (quality, validity, size, outcome type, condition)
- Reported both efficacy (NNT) and safety (local, systemic, withdrawals) outcomes
Limitations
- Trials were short (outcomes mainly at two weeks), so long-term efficacy and safety of ongoing use could not be assessed
- Many included trials were small, allowing chance to influence event rates
- Trials spanned several decades and used different preparations, formulations, concentrations and application schedules, creating clinical heterogeneity
- Reported outcomes were inconsistent, requiring a constructed hierarchy of outcomes
- Insufficient data to compare efficacy between different individual topical NSAIDs
- Possible publication bias could not be ruled out, though no additional unpublished trials were found
Key Takeaways for Patients
What This Means for You
- 01Rub-on (topical) anti-inflammatory gels, creams or sprays can meaningfully reduce chronic musculoskeletal pain such as osteoarthritis over a couple of weeks — about one in four to five people gets relief they would not have had from a placebo.
- 02In the trials, topical NSAIDs worked about as well as anti-inflammatory pills taken by mouth for osteoarthritis of the knee or fingers.
- 03Over the short two-week period studied, side effects from topical NSAIDs were no more common than with a dummy cream, usually just mild skin reactions like rash, itching or stinging.
- 04Because the trials were short and many were small, the long-term safety and benefit of using these products continuously is still uncertain — discuss ongoing use with your clinician.
Read the Full Paper
Access the complete peer-reviewed study from BMC Musculoskeletal Disorders
View Full Study