Nonsteroidal anti-inflammatory drugs (NSAIDs) are mainstays of the treatment of osteoarthritis and rheumatoid arthritis - common examples are ibuprophen, naproxen, diclofenac. They work by decreasing the manufacture of proteins called prostaglandins that are responsible for inflammatory changes in various places in the body, including the joints. Unfortunately, prostaglandins have other functions, which can involve the blood platelets, the kidneys and the stomach lining (mucosa). It is suppression of the latter function that can lead to the side effects most often associated with taking NSAIDs - dyspepsia and stomach pains. Occasionally these side effects can be serious - in one large study, about 1.5% of patients with rheumatoid arthritis taking NSAIDs for 1 year had a serious stomach complication. Although this seems quite a small proportion of patients, it represents a large number of patients because of the vast quantities of NSAIDs consumed.

Some years ago it became known that there are two types of enzymes that produce prostaglandins. These are called COX-1 (which is found in the tissues where NSAID side-effects occur - the stomach, kidneys, etc), and COX-2 (which is found in tissues where inflammation occurs). It was clear that if a drug could be found that inhibits or suppresses COX-2, it should have anti-inflammatory effects, but no NSAID-type side effects.

At least two COX-2 inhibitor drugs have already been approved for marketing. However, the promises offered by laboratory findings have not been entirely borne out in the clinic. For this reason, some people think they should really be called COX-1-sparing NSAIDs. Clinical findings have been reviewed in the editorial summarized here.

Results of two clinical studies have just been published. In the first, over 1000 patients with rheumatoid arthritis were given celecoxib (one of the new COX-2 inhibitors), naproxen or a sugar pill (placebo). Celecoxib and the classical NSAID were equally effective in reducing the signs and symptoms of the disease. Stomach symptoms occurred in 19% of the patients on placebo, 31% of those on naproxen, and 26% of those given various doses of celecoxib. Small ulcers in the stomach wall lining were seen using an endoscope in 4% of patients on placebo, 26% of those on naproxen and 5% of those given celecoxib. These ulcers are not clinically important - they don't cause symptoms - but their low numbers with celecoxib show the potential benefit of this type of anti-inflammatory drug, compared to standard NSAIDs.

The second study compared another COX-2 inhibitor - rofecoxib, some traditional NSAIDs and placebo in over 5000 patients with osteoarthritis - the results from 8 similar studies were combined to reach this large number. Again, the effectiveness of the two types of drug in treating the symptoms of arthritis was similar. Special attention was paid to complications of stomach ulcers (perforations, painful ulcers, bleeding ulcers). Over 12 months, ulcer complications were twice as common in patients given traditional NSAIDs as in those treated with rofecoxib. On the other hand, dyspepsia (indigestion pain) was reported equally by patients given rofecoxib and those taking standard NSAIDs.

These results show that the benefits of COX-2 inhibitors with regard to stomach symptoms are quite small. Also, the decrease in ulcers seen with the endoscope is not very important clinically, as these ulcers don't cause symptoms in most patients.

The question to be answered, therefore, is whether the benefits of COX-2 inhibitors are sufficiently great to justify a large increase in the cost of treatment. The authors of the editorial point out that, in the USA, the additional cost of prescribing a COX-2 inhibitor instead of a generic NSAID is roughly $750 a year. Cost-considerations, therefore, favor the use of a standard NSAID for patients with a low risk for ulcer complications - these are usually younger patients without a history of stomach trouble.

Older patients are at a higher risk for ulcers that can become problematic - they may have a previous history of ulcer with gastrointestinal bleeding, or cardiovascular disease. People over 75 who have previously had an ulcer have a 5% risk of developing an ulcer with complications when taking a standard NSAID. For such patients, the increased risk might justify the cost of switching them to one of the newer type of drugs.

It is possible that the new COX-2 inhibitor drugs will also have fewer side effects involving the kidneys or the bowels than the standard NSAIDs. However, before they can be recommended for widespread use in all arthritis patients, further clinical studies need to be completed in patients with a variety of such problems. The full side effect profile of a new drug only becomes known when a lot of patients have been treated.