At the coroner’s office in Dayton, Ohio, Dr. Mark Edgar stood over the body of Robert Van Hook. The deceased 58-year-old weighed 228 pounds; he wore blue pants, a white shirt, and identification tags around his ankles. Edgar, a professor of pathology at Emory University School of Medicine, had done countless autopsies over the years. But this would be the first time he examined the body of someone executed by the state.
Van Hook had died one day earlier, on July 18, 2018, inside the death chamber at the Southern Ohio Correctional Facility in Lucasville. After a tearful apology to his victim’s family, he was injected with 500 milligrams of midazolam — the first of a three-drug formula adopted in 2017. Media witnesses described labored breathing from Van Hook shortly afterward, including “gasping and wheezing” loud enough to be heard from the witness room. Nevertheless, compared to recent executions in Ohio, things seemed to go smoothly.
Still, Edgar had cause for concern. For the past few years he had been examining the autopsy reports of men executed using midazolam across the country. He found a disturbing pattern. A majority showed signs of pulmonary edema, an accumulation of fluid in the lungs. Several showed bloody froth that oozed from the lungs during the autopsy — evidence that the buildup had been sudden, severe, and harrowing. In a medical context, where a life is to be saved, pulmonary edema is considered an emergency — it feels like drowning. Even if someone is to be deliberately killed by the state, the Constitution is supposed to prohibit cruel and unusual punishment. To Edgar, the autopsies showed the executed men felt the panic and terror of asphyxiation before they died.
Experts likened the effect to being buried alive while feeling fire in one’s veins.
Edgar’s findings compounded a more familiar problem with midazolam. Unlike the drugs used to provide anesthesia in previous execution protocols, it was a sedative, not a barbiturate. Anesthesiologists had warned for years that even at extremely high doses, midazolam did not have the properties to render a person insensate — immune to pain — for the purpose of lethal injection. The second and third drugs in Ohio’s protocol — a paralytic that stops respiration and potassium chloride to stop the heart — were known be excruciating if injected without a proper anesthetic. Experts likened the effect to being buried alive while feeling fire in one’s veins.
The U.S. Supreme Court dismissed this evidence in the 2015 ruling Glossip v. Gross, waving forth a slew of executions using midazolam. But Edgar’s findings were new. Days before arriving in Ohio to do the autopsy on Van Hook, he had presented them publicly for the first time at a trial in Nashville, Tennessee, where lawyers were challenging plans to kill a man named Billy Ray Irick. Among the witnesses who corroborated Edgar’s conclusions was a leading pharmacologist, Dr. David Greenblatt, who did some of the earliest clinical trials of midazolam. He explained how a massive dose of the acidic drug would almost immediately begin to destroy pulmonary capillaries and lung tissues upon injection, leading to pulmonary edema. Other witnesses gave firsthand accounts of executions in which the condemned had struggled to breathe. This was compelling — if not quite exhaustive — evidence. “The ideal situation would be for me to do these autopsies myself and see exactly what I was interested in,” Edgar said on the stand in Nashville. The next day, Ohio executed Van Hook.
Ohio does not conduct autopsies following executions. But Ohio Federal Public Defender Allen Bohnert secured permission on Edgar’s behalf. “The autopsy was conducted in the usual manner,” Edgar wrote in a subsequent report. He made a Y-shaped incision into the chest and abdomen. A technician removed and weighed Van Hook’s organs and Edgar examined them, looking for anything unusual. When he got to the lungs, he found “significant abnormalities.” They were unusually heavy — one telltale sign of congestion. When he cut into them, he found a mix of blood and frothy fluid.
Of the 27 previously available autopsy reports for people executed using midazolam, Edgar had found evidence of pulmonary edema in 23. Van Hook was the 24th. A few weeks later, Tennessee used midazolam to execute Irick, who moved and made choking sounds — another grim sign. In a motion seeking a stay of execution and preliminary injunction for Warren Keith Henness, who was scheduled to die in Ohio in February 2019, Bohnert urged a federal magistrate judge to consider these recent developments. “At some point the courts cannot explain away the ever-growing mountain of evidence” against midazolam, he wrote.
Magistrate Judge Michael Merz granted an evidentiary hearing. After four days of testimony, he issued a damning 148-page order on January 14. The evidence surrounding midazolam had become far more persuasive since Merz last presided over such a proceeding. Not only was he now convinced that midazolam had no analgesic properties, but the drug was “sure or very likely” to cause pulmonary edema, which was akin to “waterboarding.” Yet Merz said he could not stop Henness’s execution. Under Glossip, people challenging lethal injection protocols had to prove that there was an alternative method readily available for the state to use to kill them. Henness had not met this burden.
“This is not a result with which the court is comfortable,” Merz wrote. “If Ohio executes Warren Henness under its present protocol, it will almost certainly subject him to severe pain and needless suffering. Reading the plain language of the Eighth Amendment, that should be enough to constitute cruel and unusual punishment.”
Ohio seemed poised to carry out Henness’s execution. But then, on January 22, the governor’s office issued an order of its own. Newly inaugurated Gov. Mike DeWine granted a warrant of reprieve, delaying Henness’s execution until September. In the meantime, he ordered a review of the state’s options and an examination of “possible alternative drugs.”
“Agony and Horror”
Among those who have fought the legal battles over lethal injection, the events in Ohio were a big deal. Courts across the country have repeatedly upheld the use of midazolam despite mounting evidence of its dangers. Just last week, a federal judge in Alabama denied a request by Domineque Ray to be executed by nitrogen hypoxia instead of the state’s midazolam-based protocol. (That execution, scheduled for tonight, is currently on hold for unrelated reasons.) Although Merz’s order did not stop Henness’s execution, his findings were blunt and unequivocal in a way that other judges have not been — which could influence future litigation.
The reprieve for Henness was also particularly notable coming from DeWine. As Ohio’s attorney general from 2011 through 2018, he spent much of the past decade fighting to push through executions, even as Ohio adopted new and untested protocols. Ever since U.S. executions first became derailed by a shortage of sodium thiopental — the fast-acting barbiturate long used for lethal injection — states across the country had been engaged in a macabre human experiment. Ohio was particularly eager to tinker with its formulas, adopting one-drug, two-drug, and three-drug protocols over time. After the state first used midazolam in the notoriously ugly 2014 execution of Dennis McGuire, DeWine was instrumental to passing legislation to conceal the identity of pharmaceutical companies that sold the state drugs for executions. But the risk of another botched execution appears to be of more concern now that DeWine is governor. When Henness’s attorney sent him a reprieve request highlighting Merz’s ruling, he acted within 24 hours.
Like other states that have used midazolam for executions, Ohio had been cautioned before killing McGuire — one anesthesiologist warned that it could cause “agony and horror.” The execution was a ghastly ordeal. Witnesses said he grunted and fought for air, “snorting, gurgling and arching his back,” as described in a lawsuit brought by McGuire’s family. According to the suit, one prison official “mouthed ‘I’m sorry’” to his relatives in the execution chamber.
For the next 3 1/2 years, Ohio did not carry out a single execution. But other states continued to use midazolam. Although protocols varied from state to state, each relied on quantities of the drug that were well above the average therapeutic dose. Witness accounts were consistent, often describing labored breathing from the condemned.
“If the individual was in any way aware of what was happening to them it would be unbearable.”
In 2017, Edgar reviewed the autopsy report of 39-year-old Ricky Gray, executed in Virginia that January. News reports had described unusual movements after the midazolam was administered. He “looked around, moved his toes and legs,” the Richmond Times-Dispatch reported. “He appeared to take a number of deep breaths and he appeared to make snoring or groaning sounds.” Gray’s family requested his autopsy report, which was obtained by The Guardian. “It notes that ‘blood-tinged fluid is present from the mouth’ and that ‘the upper airways contains foamy liquid,’” reporter Ed Pilkington wrote. “It also finds that the body’s lungs were ‘severely congested’ and that there were ‘red cells present in the airways.’” Edgar told Pilkington it was evidence of acute pulmonary edema. “When it is this severe you can experience panic and terror,” he said. “If the individual was in any way aware of what was happening to them it would be unbearable.”
But the executions continued apace. Shortly after that revelation, Virginia used the same protocol to kill 35-year-old William Morva. Witnesses said he, too, appeared to move and gasp for air. In the fall of 2017, Ohio ended its de facto moratorium with a revamped midazolam protocol, killing Ronald Phillips and then Gary Otte, both of whom seemed to struggle, according to witnesses. A third execution, that of 69-year-old Alva Campbell, was aborted after the execution team failed to find a viable vein. (He died three months later.)
A Tipping Point
On the occasions when botched executions have made national news, prison officials have often been the ones in the spotlight. Governors and attorneys general have also attracted a share of the blame. News reports have repeatedly exposed cavalier (and sometimes illegal) attempts to find drugs for executions from sketchy sources. Yet outside legal circles, there has been comparatively little critique of the courts that have enabled such executions to go forward. In his motion before Merz, Bohnert framed the issue in terms of a deeply entrenched problem that goes beyond lethal injection: junk science and the courts’ stubborn refusal to allow new evidence to change existing law.
“This court, and others, have misapplied the science involved in lethal injection challenges involving midazolam, leading to conclusions that are inaccurate or simply not true from a scientific perspective,” Bohnert wrote. Litigants “untrained in medicine” have “muddied the factual picture,” he wrote, while judges, “typically facing the urgent press of an impending execution,” have often further distorted the issue in their rulings. Because courts base their decisions on legal precedent, Bohnert wrote, “each subsequent judicial decision has calcified those flaws, further insulating them from the rigorous re-evaluation that is necessary in the wake of executions that demonstrate a consistent pattern of troubling inmate reactions.”
“This court, and others, have misapplied the science involved in lethal injection challenges involving midazolam.”
To Edgar, this made Merz’s ruling a pleasant surprise. “He independently took a look at this information anew and came to his own conclusion,” he said. Like Bohnert, Edgar invoked examples of faulty forensics that have been exposed over the years and the duty to correct these problems in the courts. “There are so many examples of junk science,” he said. “If we’re not willing to accept and relearn [based on new findings], then we’re in serious trouble.”
When it comes to the case law on lethal injection, Glossip is a major part of the problem. Like other decisions that have enshrined junk science into law, the ruling gave legal legitimacy to the findings of a pharmacist who had been widely discredited even before Supreme Court justices heard oral arguments. Critically, Bohnert points out, the ruling conflated insensateness and unconsciousness, two concepts whose distinctions are hugely consequential. This error has been replicated by the lower courts, while execution after execution has shown that unconsciousness cannot protect people from the tortuous effects of lethal injection.
Bohnert points to an irony about Glossip and its legacy. “The fact that the states have been allowed to continue to execute using midazolam is in large part what allowed the evidence to accumulate,” he says. Although it is not clear what will happen next in Ohio, “I think we had a tipping point here.”
Elsewhere, midazolam is on its way out. Florida and Arizona have discarded the drug, while Tennessee has not used it since the Irick execution, turning instead to the electric chair — at the request of the condemned. If Alabama uses midazolam to kill Domineque Ray tonight, his execution stands to be another “data point,” in Merz’s words, more evidence that states may or may not be torturing people to death while the courts continue to look away.