By the time the blinds were raised at 6:23 p.m. on April 29, 2014, to show Clayton Lockett strapped to the gurney and positioned to die, there was a lot that witnesses in Oklahoma’s death house had not seen.

They did not see how, for nearly an hour, a paramedic and physician tried and failed to insert an IV line into various parts of Lockett’s body, including his neck and feet.

They did not see how, after he was punctured some 14 to 16 times, Lockett’s pants and underwear were cut off so that the doctor could clumsily inject the IV into his femoral vein, near his groin, using a needle too small for the task. Nor did witnesses see the IV, which the warden chose to cover with a blanket to protect his genitals from view, but also in the name of “dignity.”

They did not see the makeshift rope that had been found earlier that day inside Lockett’s holding cell, or the lacerations on Lockett’s arms where he had slashed himself with a razor. Or the prison task force that came for Lockett early that morning, forcing their way into his blood-stained cell after he tried to block the door and subduing him with a TASER.

But what witnesses would see once Lockett was finally displayed before them was a human experiment — the first execution in the state using 100 milligrams of a new drug, midazolam, to kick off its three-part cocktail. It would go terribly wrong. As the drugs started flowing, and after he had already been deemed unconscious, Lockett jerked his head, and began to writhe and moan. “Oh my God,” Warden Anita Trammel later recalled thinking. “He’s coming out of this. It’s not working.” In the overflow room where others watched on a TV monitor, “It was like a horror movie,” one official told the Guardian. “He kept trying to talk.” Witnesses heard Lockett say things like, “something is wrong,” and “the drugs aren’t working” and “this shit is fucking with my mind.” After nine minutes, the blinds were hastily closed. The blanket was lifted to reveal that the drugs were seeping into the tissue of his inner thigh instead of his veins, causing his skin to swell.

Officials debated whether they should keep trying to kill Lockett or else try to save his life. They called the governor’s office. They decided to halt the proceeding. But then, just after 7 p.m., Clayton Lockett finally died.

On his death certificate: “Judicially Ordered Execution.”

FILE - This July 25, 2014 file photo shows bottles of the sedative midazolam at a hospital pharmacy in Oklahoma City. Oklahoma plans to resume executions Thursday, Jan. 15, 2015, after botching its last one and will use the same three-drug method as a Florida lethal injection scheduled for the same day. The drug mixture begins with the sedative midazolam and includes the same drugs used in Oklahoma’s botched execution of Clayton Lockett, who writhed on the gurney and moaned after he’d been declared unconscious. (AP Photo/File)

Bottles of the sedative Mizazolam. (AP)


On Wednesday, April 29, exactly one year after Lockett’s grisly death, Oklahoma Solicitor General Patrick Wyrick rose confidently before the U.S. Supreme Court to defend the way his state kills prisoners. Not that he should have to, he suggested. It is “a matter of fact” that 500 milligrams of midazolam will “with near certainty” render a person “unconscious and unable to feel pain” during a lethal injection. A district court in Oklahoma has said so. So have multiple Florida courts. It was up to prisoners to prove otherwise.

The case at hand was Glossip v. Gross, filed on behalf of four other men on Oklahoma’s death row, one of whom, Charles Warner, was executed just one week before the Court took the case. (His last words were reported as “My body is on fire.”) The petitioners were challenging the use of midazolam in lethal injections as cruel and unusual punishment. Not because of trouble with the way it is administered — even messy ordeals like Lockett’s don’t represent a “substantial risk” of tortuous or lingering death, the Court found in Baze v. Rees, which upheld lethal injection in 2008. (“An isolated mishap alone does not violate the Eighth Amendment,” Chief Justice John Roberts wrote in the 7-2 plurality opinion.) Rather, as federal public defender Robin Konrad told the justices on Wednesday, midazolam itself lacks the critical properties to “maintain the deep coma-like unconsciousness” that Baze deemed necessary for a humane lethal injection. Lockett’s death – and at least two others – have demonstrated this.

Lethal injection was originally designed to work in three parts. The first drug, traditionally a barbiturate called sodium thiopental, anesthetized the prisoner. The second, pancoronium bromide, induced paralysis. And the third, potassium chloride, stopped the heart. The first drug was particularly important: absent a proper dose of sodium thiopental, the effects of the second and third drugs would be indisputably excruciating, akin to being burned alive.

But soon after this protocol was upheld in Baze, sodium thiopental began to dry up, in part due to activist pressure on international suppliers. Death penalty states desperately sought new sources for the drug, then looked to replace it. Before long, states across the country were experimenting with protocols — and moving further and further away from the method approved by the Court. In the meantime, prisoners were being executed — sometimes quite tortuously — with drugs that were expired, adopted in secret, and acquired in violation of federal law.

Today, the three-drug protocol that prevailed in Baze is obsolete. Most states have moved on to single-drug executions using the barbiturate pentobarbital — but that drug, too, is now in short supply. Oklahoma’s protocol, borrowed from Florida, was designed, at least on the surface, to mimic the original lethal injection design  — the second two drugs remain the same. But midazolam is a benzodiazipane, not a barbiturate — an entirely different class of drug. The former is most commonly used as an anti-anxiety medication, to treat insomnia, or as a sedative during minor operations. During oral arguments, even Justices Alito and Antonin Scalia seemed to grudgingly acknowledge that midazolam is far from an ideal substitute.

That the Court again found itself discussing lethal injection at all seemed to irritate the judges. Justice Samuel Alito blamed “a guerrilla war against the death penalty.” Activists have made it “impossible for the States to obtain drugs that could be used to carry out capital punishment with little, if any, pain,” he complained. “And so the States are reduced to using drugs like this one.” Justice Scalia, too, inveighed against abolitionists for making it “impossible to get the 100 percent sure drugs,” referring to sodium thiopental and pentobarbital. “I guess I would be more inclined to find that [midolazam] was intolerable if there was even some doubt about this drug when there was a perfectly safe other drug available,” he said. In other words, the lack of good alternatives might just make midolazom good enough in his book.

But there is ample evidence to show that midazolam is far from good enough.  Never has it been used as a sole anesthetic during painful procedures. As 16 professors of pharmacology wrote in an amicus brief to the Court, “Midazolam is incapable of rendering an inmate unconscious prior to the injection of the second and third drugs in the State of Oklahoma’s lethal injection protocol.”

The state offered precious little proof to the contrary. Its own brief relied on the expertise of a certified pharmacist named Dr. Roswell Lee Evans, a man who has never administered midazolam, but who, at Oklahoma’s behest after an investigation into Lockett’s execution, calculated that 500 milligrams of the drugs should be enough to kill a person. He then “extrapolated” that logically, somewhere along the way, the drug would surely cause a person to fall into a coma, which would then shield a prisoner from pain.

“But his evidence for this was zero,” Justice Steven Breyer protested, cutting to the heart of the matter. Indeed, Dr. Evans arrived at his conclusions through a vague combination of expertise, hunches and research on websites like (That site warns it is “not intended for medical advice, diagnosis or treatment.”)

Justice Sonia Sotomayor seemed particularly disgusted by the state’s lack of evidence, at one point interrupting Solicitor General Wyrick to say that she was “substantially disturbed” by the inconsistencies she found between the state’s claims and the evidence used to support them. “Nothing you say or read to me am I going to believe, frankly,” she said, “until I see … with my own eyes the context. Okay?”

Wyrick tried to explain away the holes in his case by reiterating that it is up to the prisoners, not the state, to prove the only “constitutionally relevant” question: whether midazolam has “a ceiling effect that kicks in before we get to a level where [prisoners are] unconscious and unaware of the pain.” No one seems to know exactly where that ceiling lies. So while the state concedes that there is a possibility that midazolam will wear off mid-execution, it argues that this does not mean it definitely will. This level of uncertainty over midazolam is apparently not too high for Oklahoma to stop killing people with it.

Justice Elena Kagan found the logic galling. If it’s true that experiencing the effects of potassium chloride is “like being burned alive,” she said, then this is like telling someone, “We’re going to burn you at the stake, but before we do, we’re going to use an anesthetic of completely unknown properties and unknown effects. Maybe you won’t feel it, maybe you will. We just can’t tell.”

“That isn’t the world we live in,” Wyrick responded.

In the world outside the Court on Wednesday afternoon, just hours after oral arguments ended, the muckraking website Oklahoma Watch broke news of a letter from a drug company named Akorn, which had supplied the midozalam used to kill Clayton Lockett.  The letter was addressed to the Oklahoma Attorney General, who was in Washington that day. The date was March 4, 2015, mere weeks after the Court had agreed to review Glossip. The letter demanded that Oklahoma return its supply of midazolam “for a full refund” and included a line printed in italics and bold: “Akorn strongly objects to the use of its products in capital punishment.” Whether Oklahoma ever returned the drugs is unclear, Oklahoma Watch reported. What is clear is that the state spent the morning defending midazolam as perfectly appropriate for lethal injections, while knowing that at least one of its own suppliers disagrees. “Even the FDA label indicates that induction of anesthesia is a commonly accepted use” for midazolam, Wyrick told the court. In its letter, Akorn called using midazolam in lethal injection “clearly contradictory” to its intended, FDA-approved use.

Other states have received the same request from Akorn, including Arizona, where a man named Joseph Wood took a horrific two hours to die last summer, after being injected with 750 milligrams of midazalom. The letter to Arizona, also dated March 4, was obtained through open records requests by The Arizona Republic and reported less than a week before oral arguments in Glossip. The question before the justices, reporter Michael Kiefer wrote, “may be moot if Oklahoma and Arizona and the other two states that use midazolam can no longer procure it.”

As many states hold off on executions while they wait for a decision, it is difficult to guess which way the Court will rule. Justice Anthony Kennedy, understood to be the critical swing voter, was mostly silent during oral arguments. Regardless, any ruling in Glossip is widely assumed to have limited reach, if not a limited shelf life. Of the 13 executions so far in 2015, only two have been carried out using the midozolam three-drug protocol.

In Florida, which pioneered the use of midazolam — and which Oklahoma lauded for its “impressive track record” — a guest editorial appeared in The Tampa Tribune on Wednesday. In it, anesthesiologist Michael C. Lewis urged readers not to buy the “myth of Florida’s perfect executions,” calling attention to an aspect of lethal injection that seems to have been forgotten in the debate over midazolam.

“Although the safety and efficacy of midazolam is the subject of the U.S. Supreme Court case,” he wrote, “it’s the next drug in Florida’s protocol that concerns me most: the paralytic which prevents the prisoner from showing distress.”

Lewis called the paralytic “terrifying.” It was originally incorporated into the three-drug protocol by the inventor of lethal injection, an Oklahoma forensic pathologist named Jay Chapman, who wished to mask the ugly visual evidence associated with murder. (Now 75, Chapman told the Guardian on the eve of Glossip that he is ambivalent about the death penalty.) But the paralytic also meant that, for years, witnesses could not necessarily tell if something was going wrong with an execution. Prisoners would be frozen, even while suffering immensely. Sometimes, the evidence would only be revealed in an autopsy.

The narrative at the Supreme Court on Wednesday was that we used to have reliable lethal injection drugs. But in reality, as anesthesiologist Dr. Joel Zivot told Slate’s Dahlia Lithwick in advance of Glossip, “the paralyzer for a long time has been hiding … an important fact here. That these drugs don’t do even what the Supreme Court imagined that they do.”

It is an important point to keep in mind as the Court considers the future of lethal injection this time around. “The paralytic covers a lot of sins,” Zivot said. “A lot of sins. And when the paralytic is not there, the sins are revealed.”

Photo composite: Gurney: AP