Background Adrenaline is used increasingly in the management of septic shock, but its efficacy and safety are uncertain. Methods In an open, randomised, crossover study we compared the effects of stepped doses of adrenaline 0.1 to 0.5 mu g/kg per min and dopamine 2.5 to 10 mu g/kg per min on the haemodynamic and acid-base status of 23 patients critically ill with severe sepsis (n=10) or severe malaria (n=13). Findings All patients completed the dopamine study whereas in 16 (84%) patients the adrenaline infusion had to be terminated before reaching, or during, the maximum dose because of lactic acidosis (p<0.0002). Adrenaline was associated with a mean (95% Cl) increase in plasma lactate of 3.2 (2.6 to 3.8) mmol/L, and mean falls in arterial pH of 0.052 (0.035-0.068) pH units and base excess of 3.8 (2.8-4.7) mmol/L. The geometric mean (95% CI) lactate increment per unit adrenaline dose was 8.2 (5.8-10.5) mmol/L per mu g/kg per min. In contrast dopamine was associated with a fall in lactate of 1.0 (0.4-1.5) mmol/L, a rise in base excess of 1.4 (0.7 to 2.0) mmol/L (p<0.0001 in each case), and no effect on arterial pH. Both drugs induced significant increases in cardiac index and oxygen delivery with smaller increases in oxygen consumption and falls in systemic vascular resistance which were similar in severe malaria and severe sepsis (p>0.1 in each case), but there was no increase in oxygen consumption. Interpretation Infusion of inotropic doses of adrenaline in severe infections causes lactic acidosis.
