Nonsteroidal anti-inflammatory drugs (NSAIDs) such as ibuprofen and naproxen are used to treat osteoarthritis, but they can also have unpleasant effects on the stomach. That's why the labels say to take them with food—and it's also why the clever drug companies have come out with cox-2 inhibitors such as Celebrex and Vioxx that are supposed to give you the same pain relief without messing up your stomach. Some guidelines for treating arthritis also suggest topical NSAIDs, which are smeared on the arthritic joint instead of swallowed. A group of researchers at the University of Nottingham looked at four decades of studies on topical NSAIDs to see if they helped.

What the researchers wanted to know: Do topical nonsteroidal anti-inflammatory drugs (NSAIDs) work on osteoarthritis?

What they did: The researchers did a literature review—they found reports of trials from 1966 to the end of October 2003, that compared topical NSAIDs with placebo or oral NSAIDs. They looked only at studies of patients with osteoarthritis—so they excluded studies of rheumatoid arthritis, other joint pain, pain from surgery, and anything else that wasn't only osteoarthritis. They found 13 reports; most compared topical NSAIDs with placebo, and most used patients with arthritic knees. No study lasted longer than four weeks, and most were only two weeks long.

What they found: Although they may help for very short periods, there is no evidence that topical NSAIDs cure arthritis pain long term, according to this analysis. In the first two weeks of treatment, topical NSAIDs reduced pain more than placebo but not as much as oral NSAIDs—and in the third or fourth weeks of treatment, the topical NSAIDs were no better than placebo. Pain is only one symptom the drugs should improve; joint function and stiffness also count. But topical NSAIDs were also weak there—better than placebo for two weeks but not later. Unsurprisingly, patients using topical NSAIDs had fewer stomach problems than those on oral NSAIDs, but they had more itches and rashes.

What the study means to you: Topical NSAIDs may not be much use. It's hard to tell without more research, though.

Caveats: Oral NSAIDs and topical NSAIDs have only been compared in three trials, so it's hard to be definite about the differences between them. The authors say an ongoing study comparing oral and topical ibuprofen should give more answers. Also, the trials used topical NSAIDs with four different ingredients, some of which (salicylate) worked better than others (eltenac). This study only covered arthritis, so topical NSAIDs may work on other conditions.

Find out more: The British Medical Journal publishes a lot of interesting studies. The BMJ's editor recently retired after 25 years and left behind an essay on medical journals, which, he admits, may sound "grouchy".

Read the article: Lin, J., Zhang, W., Jones, A., and M. Doherty. "Efficacy of Topical Non-Steroidal Anti-Inflammatory Drugs in the Treatment of Osteoarthritis: Meta-analysis of Randomised Controlled Trials." British Medical Journal. Aug. 7, 2004, Vol. 329, pp. 324–326.

Abstract online: http://bmj.bmjjournals.com (The full article is free—click on "Full text of this article" or "PDF of this article" to the right).