Occult gastrointestinal blood loss induced by a new buffered aspirin preparation (Ostoprin, 4.0 g per day) was compared with that from an enteric-coated aspirin (Ecotrin, 3.9 g per day) in 40 patients with osteoarthritis. Blood loss was measured by the radiochromium method and compared with the HemoQuant assay of fecal heme and heme-derived porphyrins. By radiochromium, mean daily blood loss during the first week of treatment with Ostoprin increased 1.6 ml above basal compared to 0.8 ml above basal with Ecotrin (p = 0.06). When aspirin was ceased, blood loss returned more rapidly towards normal in the Ostoprin group (p less than 0.01). By HemoQuant, mean fecal hemeporphyrin excretion rose 0.63 mg hemoglobin equivalent per g feces during the first week of treatment by Ostoprin compared with 0.40 mg/g for Ecotrin (p = 0.06). There was a significant linear relationship between the two methods (r = 0.65, p less than 0.001). Serum salicylate levels achieved with both preparations were almost identical and encompassed the therapeutic range. HemoQuant is sufficiently sensitive to detect low-level aspirin-induced bleeding. Ostoprin appears to be a safe alternative to Ecotrin and has similar bioavailability.