Background: High dose insulin (HDI) has proven superior to glucagon and catecholamines in the treatment of Poison-Induced Cardiogenic Shock (PICS) in previous studies. Recommendations for dosing of HDI vary, from 0.5 U/kg/hr up to 10 U/kg/hr. The optimal dose of HDI in PICS has not been previously studied or established. We hypothesized a dose of 10 U/kg/hr of HDI would be superior to 1 U/kg/hr in the treatment of pigs with PICS from propranolol (P) with cardiac output (CO) as our primary outcome measure.
Methods: This was a blinded, randomized and controlled trial with 4 arms consisting of 4 pigs in each arm. The arms were as follows: saline control (C), or HDI at 1 U/kg/hr (1), 5 U/kg/hr (5), and 10 U/kg/hr (10). Pigs were anesthetized and underwent placement of a tracheostomy, Swan-Ganz catheter and arterial line. All pigs received a P bolus of 0.5 mg/kg followed by an infusion of 0.25 mg/kg/min until the point of toxicity was reached, defi ned as a 25% reduction in baseline heart rate (HR) x mean arterial pressure (MAP). At this point a 20 mL/kg bolus of saline (NS) was administered. The P infusion was reduced to 0.125 mg/kg/min and continued throughout the protocol. After the NS bolus the blinded infusion of C, 1, 5 or 10 was started and pigs were resuscitated for 6 hours or until death. CO, HR, MAP, arterial blood gas, systemic vascular resistance (SVR) and serum glucose were recorded every 10 minutes. Serum potassium (K) and lactate were recorded hourly.
Results: Survival: 2 pigs died in the C arm, 1 pig died in each of the 1 and 5 arms, and no pigs died in the 10 arm. There was a statistically significant increase in CO of 1.13 L/min in the 10 arm compared to the 1 arm over the 6 hour study period (p 0.009). Given the average nadir CO of 2.0 L/min, this represents a 57% increase in CO. A linear mixed-effects regression found a statistically significant dose by time interaction, in which CO improves an average of 0.00035 L/min for every 1 u/mg/kg of insulin per 10 minute observation (p 0.00035). Thus the 10 arm was superior to the 5 arm by 0.63 L/min of CO, and the 5 arm was superior to the 1 arm by 0.5 L/min of CO at 6 hours. There was also a statistically significant difference in dose by time interaction on MAP, HR, SVR and base excess, but not lactate or K. No differences in glucose utilization were found between the three insulin arms.
Conclusion: HDI was statistically and clinically significantly superior to placebo in this propranolol model. 10 U/kg/hr was statistically and clinically significantly superior to 5 U/kg/hr, which was statistically and clinically signifi cantly superior to 1 U/kg/hr. It may be appropriate to begin HDI therapy for Poison-Induced Cardiogenic Shock at 10 U/kg/hr.