We carried out a double-blind prospective study of the effect of a filter on the incidence of phlebitis associated with intravenous infusion in 541 patients. A total of 277 patients received infusions through intravenous sets with 0.22-micron IVEX-HP filters, and 264 received infusions without filters. Each infusion was evaluated daily for a maximum of three days. The incidence of phlebitis on Days 1, 2, and 3 of the study was 14.3, 31.1, and 27 per cent for patients receiving infusions without filters and 6.8, 9.7, and 11.3 per cent for those receiving infusions through the filters (P less than 0.001). Thus, the incidence was reduced by approximately two thirds in the patients who received infusions through the IVEX-HP filters. We conclude that infusion-related phlebitis is a pervasive problem in hospitalized patients, and that it is usually caused by microparticulate components that are present in the infusion fluids and can be removed by in-line filtration.