Spiegel BM et al. (2006) Comparing rates of dyspepsia with coxibs vs NSAID + PPI: a meta-analysis. Am J Med 119: 448.e27–448.e36

Dyspeptic symptoms are the most frequent and most resource-consuming complication of therapy with NSAIDs. As NSAIDs are the predominant medications used to treat chronic arthritis, dyspeptic symptoms are a key factor in the cost-effectiveness of arthritis therapies. Nonetheless, data on the factors that influence dyspepsia development are sparse. Cyclo-oxygenase 2 inhibitors (coxibs) confer a lower gastrointestinal risk, but higher cardiovascular risk, than NSAIDs, so Spiegel et al. conducted a meta-analysis to explore the incidence of dyspepsia during arthritis treatment with an NSAID, coxib, or an NSAID plus PPI.

The team identified 32 eligible, randomized, controlled trials that compared coxib vs NSAID; coxib vs NSAID plus PPI; or NSAID vs NSAID plus PPI in patients with chronic arthritis. Dyspepsia was defined as epigastric pain, dyspepsia, and nausea.

Treatment with a coxib, or with an NSAID plus PPI, was associated with a lower risk of dyspepsia than treatment with an NSAID alone. An NSAID plus PPI showed a greater risk reduction for dyspepsia than a coxib, in the two studies that compared these treatments directly. Indirect comparison of these treatments (possible because the mean incidence of dyspepsia in the NSAID arms of studies examining a coxib vs NSAID or NSAID plus PPI vs NSAID were similar) suggested that 11 patients required treatment with an NSAID plus PPI (or 27 with a coxib) to prevent one incident case of dyspepsia.

An NSAID plus PPI is superior to a coxib for preventing dyspeptic symptoms in chronic NSAID users. Prospective studies that compare these regimens directly are needed.