The risks of agranulocytosis and aplastic anemia in relation to analgesic drug use were evaluated in a population-based case-control study conducted in Europe and Israel. Analgesic use in the week before the onset of illness was compared between 221 cases of agranulocytosis and 1425 hospital controls. Analgesics significantly associated with agranulocytosis were dipyrone (metamizol sodium), indomethacin, and butazones (phenylbutazone and oxyphenbutazone). For dipyrone, the rate ratio estimate was 23.7 in Ulm, West Germany, West Berlin, and Barcelona, Spain, and the estimated excess risk for any exposure in a one-week period was 1.1 per million. In Israel and Budapest, Hungary, where the rate ratio estimate was 0.8, there was no evidence of excess risk. In all of the regions combined, the rate ratio estimates were 8.9 for indomethacin and 3.8 for butazones, with excess risk estimates of 0.6 and 0.2 per million, respectively. Analgesic use 29 to 180 days before admission was compared between 113 cases of aplastic anemia and 1724 controls. Indomethacin (rate ratio estimate, 12.7), diclofenac sodium (8.8), and butazones (8.7) were significantly associated with aplastic anemia, with estimated excess risks for any exposure in a five-month period of 10.1, 6.8, and 6.6 per million, respectively.