In a double-blind study, heparin sodium was compared with 0.9% sodium chloride injection for use in maintaining patency of indwelling devices for intermittent intravenous infusion. Adult patients who required intermittent intravenous devices were randomly assigned to receive 1 mL of a heparin sodium 100 units/mL flush solution or a 0.9% sodium chloride flush solution. Observations were recorded for each catheter, rather than for each patient. Patients were evaluated daily for the development of phlebitis. In the 160 patients for whom complete data on catheter patency were available, there were 307 observations (170 for the heparin group and 137 for the sodium chloride group). No significant difference in the duration of catheter patency or incidence of phlebitis was observed between the groups. A difference in the incidence of phlebitis could not be excluded with confidence, but inasmuch as there was no effect on duration of catheter patency, the clinical importance of this superficial venous phlebitis is questionable. The duration of patency was significantly greater in men than in women. The use of penicillins, cephalosporins, or clindamycin, alone or in combination, was significantly associated with the development of phlebitis for both treatment groups. No other factors were found to correlate with either the duration of catheter patency or incidence of phlebitis. The results of this study indicate that heparin offers no advantage over 0.9% sodium chloride injection in maintaining the patency of intermittent intravenous devices.