Do pain-relief gels and patches actually work?

Short answer

Yes, for fresh sprains and strains. With an anti-inflammatory gel, 7-8 in 10 people had their pain cut by more than half within a week (2-3 with a placebo gel), and only 7% of a pill's dose reached the blood. For long-term knee arthritis and back pain, the effect was small.

Do pain-relief gels and patches actually work?
'Rub it on your skin → it can't reach the muscle → it's all in your head' — is that right?
When you twist an ankle or pull a muscle, an anti-inflammatory gel or patch is often the first thing you reach for. We checked whether a drug rubbed on the skin really works, how it compares with pills, and for which kinds of pain it made a difference.
It worked for sprains and strains
7-8 in 10 people, versus 2-3 with a placebo gel
An analysis that pooled 61 trials with 8,386 people found that with diclofenac or ketoprofen gel, 7-8 in 10 people had their pain cut by more than half after 7 days. With a drug-free placebo gel, it was 2-3 in 10. The placebo gel was rubbed in the same way, so the effect was not just from rubbing.
A trial in 242 people with a sprained ankle
Pain dropped almost twice as much
In a trial, 242 people with a sprained ankle used diclofenac gel or a placebo gel for 8 days. By day 5, pain when moving had dropped about 49 points out of 100 with the gel and 25 points with the placebo. More than 70% on the gel versus 21% on placebo had their pain cut by more than half, and swelling was 1/3 as large.
How does it compare with pills?
Gels had the best balance of benefit and side effects
Comparing 207 trials with 32,959 people with fresh muscle or joint injuries (not low back), gels had the best balance of benefit and side effects, followed by anti-inflammatory pills. They cut pain about 1 point more out of 10 than placebo, and opioid painkillers were no better than anti-inflammatories.
Only 7% of the pill dose reaches the blood
So side effects were no different from a placebo gel
When 76 people applied diclofenac solution to their knees for 7.5 days, the amount reaching the blood was about 93% lower than with diclofenac pills. In trials of fresh injuries, brief mild redness where the gel went on happened in about 1 in 20 people, no different from a placebo gel.
For long-term joint pain, the gap was small
Knee arthritis: 6 in 10 versus 5 in 10
In 39 trials, mostly in knee arthritis, 6 in 10 people using diclofenac or ketoprofen for 6-12 weeks had much less pain. But 5 in 10 did with a placebo gel too. Diclofenac also caused more skin reactions: 20 in 100 people versus 5 with placebo gel.
Sudden low back pain did not respond well either
The pill did better, and using both was no better
In a trial, 198 people who came to an emergency room with low back pain were split into three groups. After 2 days, a score of how much pain got in the way of daily life (out of 24) improved 10.1 points with ibuprofen pills and 6.4 points with diclofenac gel. Using both improved it 8.7 points, no better than the pill alone.
Bottom line: it works on fresh sprains and strains
For long-term joint pain and back pain, the effect was small
For pain right after a sprained ankle or pulled muscle, the gel worked for 7-8 in 10 people and was not inferior to pills. The amount reaching the blood was 7% of a pill's. For knee arthritis and back pain, the difference from a placebo gel or pills was small.