The use of a 51Cr technique to detect gastrointestinal microbleeding associated with nonsteroidal antiinflammatory drugs.

Of techniques used to evaluate gastrointestinal (GI) bleeding, use of radiochromium (51Cr)-tagged erythrocytes is the most quantitative and scientifically acceptable method. The value of this technique as well as systematic errors possible with its use are discussed. The medical literature concerning 51Cr evaluation of GI microbleeding with naproxen therapy is critically reviewed. We suggest that future studies using this technique be parallel, randomized, double-blind, and include a 1-week placebo baseline phase for all subjects. Treatment with nonsteroidal antiinflammatory drugs (NSAIDs) should last 3 to 4 weeks. A parallel group of subjects should receive placebo throughout the study. For valid statistical analyses, randomization must achieve baseline comparability of weight, height, age, and sex in the treatment groups. Data transformations may be necessary to satisfy the assumptions of the statistical model. Following these guidelines will enable investigators to better evaluate GI microbleeding during treatment with naproxen or other NSAIDs, and, hopefully, to establish the safety profiles of these drugs.