Background: Lethal injection for execution was conceived as a comparatively humane alternative to electrocution or cyanide gas. The current protocols are based on one improvised by a medical examiner and an anesthesiologist in Oklahoma and are practiced on an ad hoc basis at the discretion of prison personnel. Each drug used, the ultrashort-acting barbiturate thiopental, the neuromuscular blocker pancuronium bromide, and the electrolyte potassium chloride, was expected to be lethal alone, while the combination was intended to produce anesthesia then death due to respiratory and cardiac arrest. We sought to determine whether the current drug regimen results in death in the manner intended.
Methods and findings: We analyzed data from two US states that release information on executions, North Carolina and California, as well as the published clinical, laboratory, and veterinary animal experience. Execution outcomes from North Carolina and California together with interspecies dosage scaling of thiopental effects suggest that in the current practice of lethal injection, thiopental might not be fatal and might be insufficient to induce surgical anesthesia for the duration of the execution. Furthermore, evidence from North Carolina, California, and Virginia indicates that potassium chloride in lethal injection does not reliably induce cardiac arrest.
Conclusions: We were able to analyze only a limited number of executions. However, our findings suggest that current lethal injection protocols may not reliably effect death through the mechanisms intended, indicating a failure of design and implementation. If thiopental and potassium chloride fail to cause anesthesia and cardiac arrest, potentially aware inmates could die through pancuronium-induced asphyxiation. Thus the conventional view of lethal injection leading to an invariably peaceful and painless death is questionable.