Lethal injection was invented in 1977 by Oklahoma state medical examiner Jay Chapman, who, based on his own experiences under anesthesia, concocted the drug cocktail from an ultrashort-acting barbiturate and a chemical paralytic. He added a heart-stopping drug to the mix to provide a painless, quick death with built-in redundancy. If one drug didn't kill the death row inmate, one of the other two would. But dosage is critical to the efficacy of lethal injection according to a new study, which found that if any of the doses are off the recipient not only feels pain, but he or she also must suffer a slow death by the asphyxiation following total paralysis.

Molecular biologist Teresa Zimmers of the University of Miami Miller School of Medicine and colleagues, including a surgeon, an anesthesiologist and a lawyer, analyzed the sparse public records of executions. Only two states provide such records: North Carolina and California, the latter of which was forced to do so by court order. In each of these states, varying dosages of sodium thiopental (a barbiturate to induce anesthesia), pancuronium bromide (a muscle relaxant that paralyzes all the muscles of the body) and potassium chloride (a salt that speeds the heart until it stops) are injected in doses designed to kill condemned inmates. Though the dosages vary by state, they do not vary by inmate—each is given the same amount of the drug whether short or tall, fat or thin.

As a result, death by lethal injection is not necessarily quick or painless, according to the study published in PLoS Medicine. In North Carolina inmates took an average of nine minutes to die (and much longer before flawed drug protocols were changed), and in California cessation of the heartbeat took from two to eight minutes after the last injection of the heart-stopping potassium chloride. "When potassium chloride was added, it didn't seem to change the time of death," Zimmers notes. "This suggests that potassium chloride may not be the agent of death."

In addition, researchers found that the amounts of thiopental used may not be sufficient to render the procedure painless, based on comparisons with veterinary data. In the veterinary realm, government and professional oversight has led to the development of strict dosage guidelines for the appropriate painless killing of animals. The dosages used in human executions are, in some cases, lower by body weight than the dosages that would kill only 50 percent of mice and from which monkeys have been able to successfully recover. "The way that thiopental is administered, it would be an unacceptably low dose if the inmate was a pig scheduled for euthanasia," Zimmers says.

And, although the dosages of potassium chloride would be considered adequate to kill animals, they do not appear to have the intended effect in humans, failing to hasten the time of death. "We are doing it successfully in animals and we're doing it successfully because they've taken a hard look at it," notes Jon Sheldon, a study co-author and criminal defense attorney in Virginia. "When you do it with animals, there is no pain. It's likely there is with people."

That pain takes the form of slow asphyxiation due to an inability to use the diaphragm muscle to breathe as a result of the pancuronium bromide. "In such case death by suffocation would occur in a paralyzed inmate fully aware of the progressive suffocation and potassium-induced sensation of burning," the researchers write.

The scientists analyzed only 41 of the 891 lethal injections that have taken place in the U.S. to date (and considerably more worldwide). But many of the remaining states' drug protocols and details of their executions remain secret. Nevertheless, researchers say the small sample indicates that the cocktail is not working as intended. "This idea that this is a painless procedure is completely wrong," Zimmers says. "It's just invisible because the person is paralyzed."

"The legal standard is you can't have unnecessary or gratuitous pain," under the Eighth Amendment of the U.S. Constitution, Sheldon adds. "It seems quite likely that a number of people are suffering pain. If a change to the protocol would be fairly simple to do, then the pain you are inflicting is clearly unnecessary."