OBJECTIVE To determine whether variations in the flow rate of epinephrine solutions administered via commonly available infusion pumps lead to significant variations in blood pressure (BP) in vivo. METHODS Prospective, randomized, crossover study with factorial design, using infusion pumps with four different operating mechanisms (pulsatile diaphragm, linear piston/syringe, cyclic piston-valve, and linear peristaltic) and three drug delivery rates (1, 5, and 10 mL/hr). METHODS Two healthy, mixed-breed dogs (12 to 16 kg). METHODS Dogs were made hypotensive with methohexital bolus and continuous infusion. BP was restored to normal with constant-dose epinephrine infusion via two pumps at each rate. METHODS Femoral mean arterial pressure (MAP) was recorded every 10 secs. Pump-flow continuity was quantitated in vitro using a digital gravimetric technique. Variations in MAP and flow continuity were expressed by the coefficient of variation; analysis of variance was used for comparisons. RESULTS The mean coefficients of variations for MAP varied from 3.8 +/- 3.1% (linear piston/syringe) to 6.1 +/- 6.6% (linear peristaltic), and from 3.4 +/- 2.2% (10 mL/hr) to 7.9 +/- 6.6% (1 mL/hr). The coefficients of variation for in vitro flow continuity ranged from 9 +/- 8% (linear piston-syringe) to 250 +/- 162% (pulsatile diaphragm), and from 35 +/- 44% (10 mL/hr) to 138 +/- 196% (1 mL/hr). Both the type of pump and infusion rate significantly (p < .001) influenced variation in drug delivery rate. The 1 mL/hr infusion rate significantly (p < .01) influenced MAP variation. Cyclic fluctuations in MAP of < or = 30 mm Hg were observed using the pulsatile diaphragm pump at 1 mL/hr. CONCLUSIONS Factors inherent in the operating mechanisms of infusion pumps may result in clinically important hemodynamic fluctuations when administering a concentrated short-acting vasoactive medication at slow infusion rates.